ANALISIS TRIWULANAN: Perkembangan Moneter, Perbankan dan Sistem Pembayaran, Triwulan II - 2007
BULLETIN OF MONETARY ECONOMICS AND BANKING
Center for Central Banking Research and Education
Bank Indonesia
Patron
Dewan Gubernur Bank Indonesia
Board of Editor
Prof. Dr. Anwar Nasution
Prof. Dr. Miranda S. Goeltom
Prof. Dr. Insukindro
Prof. Dr. Iwan Jaya Azis
Prof. Iftekhar Hasan
Prof. Dr. Masaaki Komatsu
Dr. M. Syamsuddin
Dr. Perry Warjiyo
Dr. Iskandar Simorangkir
Dr. Solikin M. Juhro
Dr. Haris Munandar
Dr. Andi M. Alfian Parewangi
Dr. M. Edhie Purnawan
Dr. Burhanuddin Abdullah
Editorial Chairman
Dr. Perry Warjiyo
Dr. Iskandar Simorangkir
Managing Editor
Dr. Andi M. Alfian Parewangi
Secretariat
Rita Krisdiana, Skom., ME
Wahyu Yuwana Hidayat, SE., MA
Tri Subandoro, SE
Aliyah Farwah, SP., MSEI
The Bulletin of Monetary Economics and Banking (BEMP) is a quarterly accredited journal
published by Center for Central Banking Research and Education,Bank Indonesia. The views
expressed in this publication are those of the author(s) and do not necessarily reflect those
of Bank Indonesia.
We invite academician and practitioners to write on this journal. Please submit your paper
and send it via mail: [email protected]. See the writing guidance on the back of this
book.
This journal is published on; January – April – August – October. The digital version including
all back issues are available online;please visit our link:http://www.bi.go.id/web/id/publikasi/
jurnal+Ekonomi/.If you are interested to subscribe for printed version, please contact our
distribution department: Statistic Disemination and Management Intern Division - Department
of Statistic, Bank Indonesia, Sjafruddin Prawiranegara Building, 2nd Floor – Jl. M.H. Thamrin
No. 2 Central Jakarta, Indonesia, phone (021) 2981-6571/2981-5856, fax. (021) 3501912,
email: [email protected].
1
BULLETIN of monetary Economics
and banking
Volume 16, Number 3, January 2014
QUARTERLY ANALYSIS : The Development of Monetary, Banking, and Payment System,
Quarter IV - 2013
179
The Quarterly Report Team, Bank Indonesia
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
Durmus Özdemir, Harald Schmidbauer
183
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
M. Noor Nugroho, Ibrahim, Tri Winarno, Meily Ika Permata
205
Fund Management And the Liquidity of The Bank
Gantiah Wuryandani, Ramlan Ginting, Dudy Iskandar, Zulkarnain Sitompul
231
The Dynamics Of Total Factor Productivity of medium and Large Manufacturing
In Indonesia
Ndari Surjaningsih, Bayu Panji Permono
259
QUARTERLY ANALYSIS The Development of Monetary, Banking, and Payment System, Quarter IV – 2013
179
QUARTERLY ANALYSIS
The Development of Monetary,
Banking, and Payment System,
Quarter IV – 2013
The Quarterly Report Team, Bank Indonesia
The economic growth of Indonesian economy in quarter IV 2013 is relatively better than
Bank Indonesia’s prediction, and is with more balance growth structure. The growth increased
from 5.63% (yoy) in quarter III to 5.72% (yoy) in of quarter IV 2013. The growth is supported
by the improving real export along with the increase of demand from trading partners. On the
other side, domestic demand increase moderately reflected from the slowing down household
consumption and investment, particularly non-building investment. Overall, the economic
growth of Indonesia reached 5.78% in 2013.
The increasing export led to significant decrease of current account deficit and sustained
the improving performance of Indonesia Balance of Payment in quarter IV 2013. The current
account deficit of quarter IV 2013 decreased significantly to 1.98% of the GDP, which is
much lower than the current account deficit in the quarter III 2013 of 3.85%. Two major
contributions for export were from export in manufacture along with the increasing demand
from the United States and Japan, and export of natural resources regarding the anticipation
of export banning for mineral resources and coal (UU Minerba)1. On the other hand, the deficit
reduction on current account was also influenced by import reduction due to moderate growth
of domestic demand.
The improvement of Indonesia’s current account deficit in quarter IV 2013 was also
supported by surplus on financial capital transaction, generated from corporate foreign debt,
domestic with drawal from savings abroad, and foreign direct investment. Bank Indonesia
predicted that the strengthening Indonesia’s current account will continue to 2014, considering
the declining current account deficit along with the increasing surplus of capital and financial
transaction. On December 2013, Indonesia’s foreign exchange reserve increased to USD 99.4
billion or equivalent to 5.4 months of importand the installment of the government foreign
debt, which is above the international adequacy standard of 3 months import.
Depreciation pressure on Rupiah was lower inquarter IV 2013. This is influenced by
the improving performance of Indonesia’s Balance of Payment inquarter IV 2013 and several
policies taken by Bank Indonesia and the government. The Rupiah exchange rate, based on,
1 Act No. 4, 2009 on Mineral and Coal (Undang-undang No. 4 tahun 2009 tentang Pertambangan Mineral dan Batubara).
180
Bulletin of Monetary, Economics and Banking, January 2014
depreciated by 4.85% (qtq), much lower than depreciation in quarter III 2013 by 14.29% (qtq).
On average, Rupiah depreciate by 8.76%, slightly increased from 8.18% in quarter III 2013.
Along with the lower pressure, the volatility of Rupiah’s was also lower.
For the whole year, the rate of rupiah tended to depreciate in 2013. On average,
Rupiah depreciated by 10.4% from Rp9,358 per USD in 2012, to Rp10,445 per USD in 2013.
Nevertheless, the situation was manageable with lower volatility compared to other Asian
countries.
Responses on policies by Bank Indonesia and coordination with the Government performed
positive impact on the reduction of inflation pressure in quarter IV 2013. After the period of
high inflation pressure led by food volatility and the increase of fuel price in quarter II and III
2013, CPI inflation in quarter IV 2013 decrease from 4.08% (qtq) in quarter III 2013 to 0.75%
(qtq) in the quarter IV 2013; or equivalently from 8.40% to 8.38% (yoy).
Major contributions for the lower inflation pressure above were the deflation on some
volatile food groups and lower inflation on administered price group following the increase of
fuel price subsidy. Deflation on volatile food by 0.58% (qtq) in quarter IV 2013 was supported
by the increase of supply following the harvest period on some commodities such as red onion
and various chilies. Moreover, the relaxation of import regulation also contributed to the
availability of the commodity such as garlic.
On the other hand, administered price inflation of 1.40% (qtq) was relatively lower and
decreased sharply from 8.94% (qtq) in quarter III 2013. The sharp decrease is in accordance
with the less price regulation over strategic goods after the government increases their subsidy
on fuel price in quarter II 2013. The only inflation pressure within administered price group in
quarter IV 2013, was tariff increase on electric power stage IV in November 2013 along with
the increase of the fuel price for household, particularly the liquid natural gas (LPG) related to
the adjustment of distribution tariff at the end of the year.
Along with the two groups above, the core inflation decreased by 2.59% (qtq) to
1.00% (qtq).This is the subsequent impact of lower depreciation pressure on inflation, the
improving expected inflation after the increase of subsidized fuel price, as well as the decrease
of international price on non-food commodities, particularly gold.
The adjustment of Indonesia’s economy was well managed, sustained by the strong and
the stability of financial system. The banking industry resiliency remained solid with tolerable
credit risk, good liquidity and market condition, as well as the strong capital resiliency. The
growth of banking loan decreased from 21.9% on November 2013 to 21.4% on December
2013 (or equivalent to 17.4% by neutralizing the exchange rate depreciation). This is supported
by the slowing domestic demand and an increase of interest rate. Bank Indonesia coordinated
with OJK (Financial Service Authority) to direct the loan growth in accordance with moderate
growth of domestic demand. Meanwhile, the performance of stock market improved in 2014 as
QUARTERLY ANALYSIS The Development of Monetary, Banking, and Payment System, Quarter IV – 2013
181
indicated by the increase of Stock Price Index (IHSG). In contrast, the performance of government
bonds market weakened as indicated on the higher yield on government bond (SBN).
Moderate growth of domestic economy in 2013 reduced the payment system transaction.
The transaction value of the payment system decreased by Rp1,930 trillion (-5.31%) to
Rp34,419.79 trillion, relative to quarter III 2013. The reduction was mainly occurred on Bank
Indonesia Real Time Gross Settlement (BI-RTGS) transaction; due to lower transaction for
monetary operation. Nevertheless, the transaction volume in quarter IV 2013 increased by 94.7
million transactions or increased 9.36% compared to quarter III 2013. The increasing volume
of transaction applied for all types of payment system with the highest increase in Card based
Payment Instrument (APMK), particularly ATM card and debit/ATM card, which is commonly
used by public during holidays.
182
Bulletin of Monetary, Economics and Banking, January 2014
Halaman ini sengaja dikosongkan
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
183
INTEREST RATE RISK IN TURKISH FINANCIAL
MARKETS ACROSS DIFFERENT TIME PERIODS
Durmus Özdemir1
Harald Schmidbauer2
Abstract
A measuring the risk associated with interest rates is important since it is beneficial in taking measures
before negative effects can take place in an economy. We obtain a risk measure for interest rates by
fitting the generalized Pareto distribution (GPD) to positive extreme day-to-day changes of the interest
rate, using data from the Istanbul Stock Exchange (ISE) Second Hand Bond Market, namely Government
Bond interest rate closing quotations, for the time period 2001 through 2009. Although the use of the
GPD in the context of absolute interest rates is well documented in literature, our approach is different
insofar and contributes to the literature as changes in interest rates constitute the target of our analysis,
reflecting the idea that risk arises from abrupt changes in interest rate rather than in interest rate levels
themselves. Our study clearly shows that the GPD, when applied to interest rate changes, provides a good
tool for interest rate risk assessment, and permit a period-specific risk evaluation.
Keyword: Interest rate risk; covered interest parity; Turkey; generalized Pareto distribution
JEL Classification: G1; C1
1 Corresponding author: Istanbul Bilgi University, Department of Economics, Dolapdere Campus, Kurtulus Deresi Cad., Yahya Köprüsü
Sok.No: 1, 34440 Beyoglu, Istanbul, Turkey, Tel.+902123115326, Fax+902122970134, E-mail: [email protected].
2 Department of Business Administration, Bilgi University, Santral Campus, Eski Silahtaraga Elektrik Santrali, Kazım Karabekir Cad.
No: 2/13, 34060 Eyup, Istanbul, Turkey, Tel.+902123117789, E-mail: [email protected].
184
Bulletin of Monetary, Economics and Banking, January 2014
I. INTRODUCTION
The risk expresses the chance of occurrence of an undesired event or events and nonaccrual of an intended and/or planned expectation. In an economic sense risk is the probability
of a monetary loss regarded with a transaction or loss resulting due to decreasing financial
returns. Cyclical fluctuations and price changes can increase the risk of occurrence of the
undesired situations.
Risk is divided into two as systemic and systematic risks. All securities in financial markets
are subject to systematic risks, and systematic risks arise for example when fluctuations within
political and economic conditions affect the behavior of assets in financial markets. As a result
systematic risks are unavoidable in the sense that keeping them under control in a way is
impossible. Systemic risks on the other hand are risks related with controllable processes such
as intra-firm investment risks or a risk that may be likely to occur due to a decision on a financial
issue (Turanlı, Özden and Demirhan (2002)).
Interest rate risk should therefore be considered within the context of systematic risks.
The fluctuations in interest rates could not totally be controlled but some measures may be
taken or certain tools may be applied to cope with interest rate risk.
Measuring the interest rate risk is important since it may be beneficial in taking measures
before negative effects can take place in an economy (see Woodford (1999)). From the
perspective of finance interest rate should be considered not only with economy but with
many other factors as well. According to Ang and Bekaert (2002), risk hidden in the behavior
of interest rates has a direct effect on the functioning of markets. Duffie and Kan (1996) and
Dai and Singleton (2002) show that interest rates not only affect the functioning of markets
but also have the power to alter the structure of the markets.
There are many other perspectives as well. For example financial income perspective says
that the income going to be generated in the future is effected by interest rates because today’s
value calculation is made by an assumed interest rate level. If there is an unexpected change in
the interest rates there is a risk that the value of income may be lower than expected. From an
institutional perspective, changes in interest rates affect a financial institution’s market value
(Carneiro and Sherris (2008)). Because the value of a financial institution’s assets and liabilities
on the one hand and off-balance-sheet contracts written on interest rates on the other are
affected by a change in rates, the present value of future cash flows and in some cases even
the cash flows themselves can change.
The focal point of the present study is to measure the interest rate risk in the Turkish spot
market for government bonds. We will first look at what has happened in the Turkish economy
within the period under investigation (2001– 2009). After this we will look at the statistical
properties of changes in the daily series of interest rates.
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
185
Our analysis is based on the tail index of the generalized Pareto distribution (GPD),
applied to threshold excesses of changes in interest rate series. The tail index characterizes
whether the underlying distribution has heavy tails. A similar approach (however, applied to
interest rate figures themselves, not to changes of rates) was used by Bali and Neftci (2001) in
order to compute a VaR for interest rates in the American market. Meyfredi (2005) has used
the estimation of risk measures associated with fat tails for stock market returns in several
countries. The behaviour of joint threshold exceedances of returns on international stock
indices is investigated by Schmidbauer and Rösch (2004). They show how a bivariate GPD can
contribute to financial risk assessment among markets in bull and bear periods.
Gencay, Selçuk and Ulugulyağcı (2002) applied this setting to data from Istanbul Stock
Exchange and derive a VaR measure meant to be an alert system for the market. Gencay and
Selcuk (2001) had already applied a similar methodology for overnight interest rates of Turkish
money markets in order to derive a measure querying whether the ex-ante interest overnight
levels are indicators of the 2001 crisis or not.
Next section outline the literatures study and the underlying theory on interest rate. Section
3 discuss the data and methodology, while section 4 provide result and analysis. Conclusion
will be provided on last section and concludes the study.
II. THEORY
Many factors may determine the shape of interest rate distribution. From micro level
behaviour, macro aggregated, and external factors including negative shock may change the
dynamics of interest rate; gradually or drastically. On international perspective, the rule of covered
interest provides basic relationship among interest rate, exchange rate, and inflation.
Common dynamic of interest rate is well recognized, one of them is that the interest
rate volatilities is stochastic. It is also recognized that interest rate tends to cluster, particularly
when shifting form low to high volatility; see Andersenand Lund (1997) in Allan Sall Tang
Andersen (2011). Borodin and Strokov (2011), investigates the interrelations between the interest rates and
international trade within the BRIC countries, and found that countries with lower interest rates
experience growth of the share of machinery industry exports rather than agriculture and food
products, and, on the contrary. On the other hand, in countries with higher interest rates, the
share of agriculture and food exports increases and the share of machinery industry products
declines. The investigation has shown that a relative shift in the interest rate can affect the
specialization of countries.
David Andolfatto (2012) used simple neoclassical model and show that liquidity shock at
home and foreign potentially contribute to trade imbalances and push down the interest rate,
which is claimed to be inline with Bernanke’s global saving glut hypothesis.
186
Bulletin of Monetary, Economics and Banking, January 2014
Turner (2014) noticed many advanced countries including the FED, Bank of Japan and
the Bank of England purchase government bonds on a massive scale to lower the long-term
interest rate, and to stimulate aggregate demand. It shows low long-term rate has recently
become an important intermediate target of central banks in the advanced economies, which
affect the short term rate as well. The relationship between the exchange rate, short term rate,
and long term interest rate may be illustrated as follows:
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
In a more formal formulation, Conway and Orr (2002) construct a global interest rate
model (GIRM) based on the concept of efficient debt markets, where by any arbitrage opportunity
is exploited by global investors. This model indentify three categories of the determinants of
interest rate dynamics; economic fundamentals, short run deviations of bond yields from their
long run levels, and global financial integration.
Desroches and Francis (2007) identify the behaviour of the world real interest rate
which is determined by a number of key variables that change relatively slowly over time. They
include the growth of labour force, which affects investment demand, and the age structure
of the world economy, which influences savings, and also the level of financial development.
Recently, Abbritti, Salvatore, Moreno, and Sola (2013) use FAVAR model to model the global
term structure. Using panel of international yield curves, they show that global factors account
for more than 80 percent of term premia in advanced economies, while domestic factors are
more relevant to explain the short-run dynamics of the rate.
Those factors explained above will determine the shape of interest rate, which is appearing
to have a fat tail distribtution. Fisher and Tippett (1928) originally introduced the traditional
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
187
method for modeling extreme-value data is based on the extreme-value limiting distributions.
Pickands (1975) introduced the Generalized Pareto Distribution (GPD) as a two parameter family
of distributions or exceedances over a threshold. Later on there other studies extended the theory
such as the article by Hosking and Wallis (1987). Use of GPD in economics is mainly done in
last decade. Bali and Neftci (2001), Gencay and Selcuk (2004), Gencay, Selçuk and Ulugulyagcı
(2002), Schmidbauer and Rösch (2004) and Meyfredi (2005) are to name a few.
Carr and Wu (2007) show that currency options have time-varying skewness. By using
model-free estimates of the volatility and skewness priced in interest-rate options, it can be
shown that interest rate distributions also show time-varying skewness (see Trolle and Schwartz
(2010)). The main purpose of the paper is to provide a consistent framework for modeling the
stochastic volatility and skewness. Finally, calibrating the model to time-series of the market
data is interesting, as it shows the applicability of the model.
III. METHODOLOGY
3.1. The Generalized Pareto Distribution (GPD)
Let (it) designate a daily series of interest rates (t indicates the day), and define the series
of daily changes as
‫ݎ‬௧ ൌ
݅௧ െ ݅௧ିଵ
Ǥ ͳͲͲΨ
݅௧ିଵ
(1)
The focal point of this paper is to study the upper threshold exceedance behaviour of
the series (rt) on the basis of the generalized Pareto distribution (GPD), which is a model for
excesses of a random variable. The rationale behind using the GPD is a limit theorem which
states3: Let r1,..., rn be iid random variables, and let R be distributed like ri. Then, for large n
and u, there are ξ and σ such that the distribution function of the excess (r – u) conditional on
R > u, is approximately given by:
ሺǢ Ɍǡ ɐሻ ൌ
ିଵȀஞ
‫ۓ‬
ۖͳ െ ሺͳ ൅ Ɍ ɐ ሻ

‫۔‬
ሻ
ۖ ͳ െ ሺെ
ɐ
‫ە‬
3 For example, see Coles (2001).
ۗ
Ǣ Ɍ ് Ͳۖ
ۘ
Ǣ Ɍ ൌ Ͳۖ
ۙ
(2)
188
Bulletin of Monetary, Economics and Banking, January 2014
Here, σ > 0 is a scale parameter; it depends on the threshold and on the probability
density function of ri. The shape parameter ξ is called the tail index, since it characterizes the
tail of the density function:
• The case ξ > 0 corresponds to fat-tailed distributions; in this case, the GPD reduces to the
Pareto distribution.
• The case ξ = 0 corresponds to thin-tailed distributions; the GPD then reduces to the
exponential distribution with mean σ.
• The case ξ < 0 corresponds to distributions with no tail (i.e. finite distributions).
• When ξ = 1, the GPD becomes a uniform distribution on the interval [0, σ ].
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
Figure 1. Fitting the GPD to data
A typical example of fitting the GPD to the upper tail of one of the (r_t) series under
consideration is shown in Figure 1. The histogram represents the upper tail of the empirical
distribution of daily changes in the series interest rate (456 daily closing quotations of interest
rates to maturity government bonds trading at the ISE Bounded Bond Purchasing Market;
faiz456) during period 2), where we used the 80% quintile as cutoff point. This quintile was
used as cut off point through out our study. The red line is the density of the normal distribution
with the same mean and variance as faiz456 in period 2, and the green line is the density of the
GPD fitted to the data. It is obvious that the normal distribution overestimates the probability
of moderate changes and underestimates the probability of large changes. This makes it
inappropriate for risk analysis in our case.
Computationally, we use the package “evd” (see Stephenson, 2002) within the statistic
software environment R (The R Core Team, 2013) tofit the GPD to data.The estimation method
implemented in “evd” is maximum likelihood. Standard errors were cross-checked using
bootstrap to ensure the reliability of results.
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
189
3.2. Data
We use daily closing quotations of interest rates of 90, 182, 273, 365 and 456 days to
maturity government bonds trading at the ISE Bounded Bond Purchasing Market. The periode of
analysis covers January 2001 to December 2009. This data is available upon request from ISE.
Prior the application of GPD method, we identify the presence of structural break over the
observation horizon. The investigation provides us four periods. A plot of the series is shown in
Figure 2 for the four periods under investigation. There are no corporate bonds in this market.
The Turkish Bond Market is dominated by Treasury Bonds.
IV. RESULT AND ANALYSIS
4.1. Structural Breaks in the Interest Rate Series
We shall now approach the question of how to divide the period under consideration
into sub-periods by applying a statistical test for structural changes to the time series of daily
interest rates. This will provide further arguments for a separate risk analysis in the three subperiods.4 In addition, we will clearly see the limitations of regression models when applied to
the interest rate series.
The method we use will find breakpoints in a regression relationship, with interest rates
as dependent variable and time (i.e. day) as independent variable. This method is based on Bai
and Peron (2003); its implementation is described in Zeileis, Kleiber, Kramer and Hornik (2003).
Breakpoints are computed with the objective of minimizing the residual sum of squares under
the constraint that no segment should be shorter than 15% of total time period considered.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Figure 2. Breakpoint analysis of faiz0915
4 We analyzed the period January 2001 through August 2008, based on structural breaks. The subsequent period, here called Period
4, was adjoined for economic reasons.
5 Faiz091 is a daily closing quotations of interest rates of 90 days to maturity government bonds trading at the ISE Bounded Bond
Purchasing Market.
190
Bulletin of Monetary, Economics and Banking, January 2014
(Our time series, beginning with January 2001 and ending in August 2008, are 1930 days long).
The number of breakpoints is not predetermined, but results from the procedure.
The test for structural changes finds four breakpoints in the series faiz091, which we chose
for this purpose to represent interest rate evolution. The results of the breakpoint analysis are
displayed in Figure 3. In our subsequent analysis, we shall ignore the first breakpoint and form
period 1 with 2003-10-06 as last day. This is justified because of the relative homogeneity of
circumstances and events in this period. We are there fore led to a definition of sub-periods
and their characterization as shown in Table 1.
4.2. Characteristics of the Sub-Periods
First of all, it is justifiable to separate this whole period into only two periods: the period
until 2002, and the period from 2003 through 2009. Starting from the beginning of 2001
and ending with the end of 2002 there were three events that mainly shaped this period: the
economic crises experienced on 28 February 2001, the September 11, 2001; and the Turkish
General Elections in November 2002. The period was comprised of many instabilities in terms
of both economy and politics throughout the period (see Insel, 2003).
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Between 2003 and 2008, a growth of 7% growth was seen in the Turkish economy on
the average. Per capita GDP had increased by 30%, the domestic currency has revalued 30% as
well. On the other hand a 100% set back was seen on trade and balance of payments deﬁcit.
Inﬂation dropped to 12% from 40%, while the interest rate level dropped to a figure of 21%
from a rate of 76% at the end of 20016.
6 All quoted figures are taken from: Banking Regulation and Supervision Agency (BDDK) Financial Markets Report, March-June 2006,
Number 1-2. Available online at http://www.bddk.org.tr/english/Reports/Financial Markets Report/1971fprMart Haziran2006ingilizce.
pdf. Accessed October 2008.
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
191
The period between January 2001 and September 2003
As mentioned above this period was stricken with economic and political instabilities. The
resolution which authorized the Turkish National Assembly to send troops to Iraq was approved
with a 50% majority on 2003-10-06. According to the news expressed the day after this was
perceived as a manifestation of “political integrity” by the markets7.
It should also be mentioned that the inflation was reported to be the lowest in 30 years
in October 20038. Shortly afterwards the Treasury explained a debt structuring in the sense of
swapping the short term government bonds with longer maturities. Interest rates had dropped
200 basis points, and the Turkish Government was subsequently able to borrow for longer
term.9
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
Figure 3. The faiz series, Perıod 1
The Period between October and May 2006
There were four main events 2003 shaping this period: WTO abolished trade barriers,
capital flows rendered more liberalised, growth of developed economies had increased, and
inflation in developed countries.
It can be said that this period was a period of capital flows between diverse markets.
Total volume of capital circulation throughout the world had reached approximately $15 trillion
7 Hurriyet Online: “TezkereGeçti Asker IrakaGidiyor, Kabul 358 Red 183”, date: 2003-10-07. Available online at http://webarsiv.
hurriyet.com.tr/2003/10/07/hurriyetim.asp. Accessed October 2008.
8 Hurriyet Online: “Enflasyona Eylül Çelmesi”, date: 2003-10-04. Available online at http://webarsiv.hurriyet.com.tr/2003/10/03/
hurriyetim.asp. Accessed October 2008.
9 Hurriyet Online: “Para KuruluToplandı”, date: 2003-10-15. Available online at http://webarsiv.hurriyet.com.tr/2003/10/15/hurriyetim.
asp. Accessed October 2008.
192
Bulletin of Monetary, Economics and Banking, January 2014
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Figure 4. The faiz series, Period 2
according to the IMF Economic Outlook10. Developing countries in this sense were also among
the beneficiaries. An amount of $2 trillion out of $15 trillion had flown to them, Turkey’s share
being $90 by foreign investment, according to the Turkish Central Bank Inflation Report11.
The Period between 2006-06-02 and 2008-08-29
There were four main events that shaped the period12: inflation fear of developed countries,
the increase in interest rates, the sub-prime crises through the end of the year 2007, and the
banking crises throughout the world.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
�
��
��
��
��
��
��
Figure 5. The faiz series, Period 3 and 4
10 International Monetary Fund (IMF) World Economic Outlook, October 2006, pp. 1–6. Available online at http://www.imf.org/
external/pubind.htm.Accessed October 2008.
11 Turkish Central Bank, Inflation Report 2006-IV, pp. 41–46. Available online at http://www.tcmb.gov.tr/.Accessed October 2008.
12 International Monetary Fund (IMF) World Economic Outlook, October 2008: Financial Stress, Downturns and Recoveries, pp. 1–46.
Available online at http://www.imf.org/external/pubs/ft/weo/2008/02/pdf/text.pdf. Accessed October 2008.
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
193
4.3. Statistical Properties of Daily Interest Rate Changes
Tables 4 and 5 in the Appendix give an analysis of the distributional properties of the
percent point changes in the five series for the four periods in terms of mean, variance and
standard deviation, skewness, kurtosis, minimum, median, and maximum. Our goal in the
present paper is an evaluation of the interest rate risk. Therefore, the two most important items
in the list are the variance and the kurtosis.
There are obvious differences between the periods: The range of daily changes is widest
for period 1; the variance and the kurtosis are largest for period 1. The behaviour of the five
series within the periods gives insight into the characteristics of the different maturities, but
also reveals further differences between the periods. In particular, some of the characteristics
resulting from Tables 4 and 5 are:
• The arithmetic mean of the daily changes in the faiz series increases from faiz091 through
faiz456 most pronouncedly in period 1. An explanation may be that period 1 was regarded
as risky by many investors in the sense that the Turkish financial market’s risk premium is still
high. As a consequence, investors demanded high long-maturity interest rates as a compen
sation for risks in future periods.
• The variance increases from faiz091 through faiz456 through out all periods, in other words:
The interest rate risk increases with maturity.
• The tail behaviour of the distributions, as expressed in the kurtosis, is more complex. The
kurtosis becomes larger as maturity increases only in period 1. This points again to an
elevated risk for higher maturities in period 1. The results of Tables 4 and 5 point to a high
risk in period 1, lower (and similar) risks in periods 2 and 3, and somewhat higher risk (albeit
reduced “surprise potential”, indicated by the lower kurtosis) in period 4. The kurtosis
generally points to heavy tails in all periods across all series, with a few exceptions. The more
complex kurtosis structure justifies using the GPD as a means to study the tail behaviour of
the interest rate change distributions.
• The ratio between minimum and maximum percentage point change is increasing with
maturity during periods 1, 2, 3, with a reduced rate during period 4. This is also clearly visible
in the boxplots in Figure 4.
• The days when minima and maxima occurred is always the same or very close in period 1.
4.4. GPD-Based Interest Rate Risk Measurement
The estimation results are reported in Table 2. In our context of risk measurement, the
estimated tail index ξˆ is more important than σˆ. As stated above, a positive tail index indicates
that the distribution of interest rate changes has a heavy upper tail (see Table 2).
194
Bulletin of Monetary, Economics and Banking, January 2014
• Estimates of the GPD parameters ξ and σ, together with their standard errors, based on
daily interest rate changes (computed as rt = ln (it - it-1))) are above their empirical 80%
quantile (that is, based on threshold exceedances of the 80% quantile) for each period,
• 95% and 99% quantiles of the interest rate changes (the columns designated as q95 and
q99, respectively),
• The corresponding quantiles are obtained by adding a GPD-based quantile to the empirical
80% quantile (which served as the threshold).
The relatively close agreement between the latter pairs throughout the periods we
considered and across the interest rate maturities can be seen as a confirmation of the model
accuracy.
A comparison of the four periods with respect to the tail properties of the interest rate
change distribution leads to the following remarks:
• Period 1 has exceptionally high values of ξ for each interest rate series considered: All five
tail indices are significantly positive (which indicates heavy tails, here: an elevated risk that
tomorrow’s interest rate is much higher than today’s) at the 5% level of significance.
• There is little difference between Periods 2 and 3, as far as the tail index is concerned. None
of the interest rate change distributions is heavy-tailed, with the exception of faiz456. This
points to an elevated overnight increase in interest rate only for long-term bonds.
• The exceptional status of faiz456 has disappeared in Period 4.
• The normal distribution is not appropriate to measure the risk associated with interest rates
in Turkey. The GPD, derived as an explicit model for distribution tails, fits very well and
provides a close fit between the theoretical VaRs and empirical quantiles.
4.5. Discussion
In this section we will try to discuss the economic implications of our study to explain
how our study fits into economic arguments. We have examined the interest rate risk in the
Turkish economy. Statistical modelling is the key to the development of such risk scores.
There are many economic reasons to have different levels of risks in interest rates. Statistical
risk scores can make a useful contribution to economic decision making under uncertainty. A
general assumption of a decision or action depends on the occurrence of some adverse event,
and that we have data which indicate how the likelihood of this event depends on values of
observable risk factors.
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
Period 1
Period 2
Period 3
Period 4
195
Figure 6. Boxplots of Interest Rate Changes, Four Periods
A risk indicator can be provided by the examination of extreme values of interest rates
in the framework of extreme value theory. Extreme value theory is a powerful framework to
study the tail behaviour of a distribution; hence it can be a good indicator for risk.
196
Bulletin of Monetary, Economics and Banking, January 2014
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
For this study we use the generalized Pareto distribution (GPD) to asses the interest rate
risk for the period starting in 2001 till the end of 2009. Estimating GPDs to the data resulted
in a good fit between the model and our data for all periods and maturities. It turned out
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
197
that the tail indices, indicating the weight of the upper tail of distributions of daily interest
rate changes, became smaller and smaller, indicating that tails became thinner from period to
period (except for the maturities second period faiz456 and faiz456, faiz356 on period3), thus
reducing interest rate risk.
As summarized in Figure 3, the interest rates have three significant structural brakes in
this period. Due to justifiable economic reasons, the entire period is divided into four sub-cycles.
These cycles are important because, for example, during recession consumers are likely to cut
back on luxury items, and thus firms in the consumer durable goods sector should see their
credit risk increased. Moreover, there is considerable evidence that macroeconomic conditions
impact the emergence of risk. For example, during economic crises reduced (or even negative)
growth will slow down the adjustment speed of capital. Economic agents borrow less. Thus
GDP growth is positively associated with the likelihood of debt issue (see Hackbarth, Miao and
Morellec (2006)).
Keeping this argument in mind, in our sample periods, the highest risk is in the first period
with a highest value of ξ. As far as high risk is concerned, the first period is followed by the
third period, but the only risky periods of borrowings are the longer maturity borrowings, i.e.
a year or more maturity borrowings. The second period, starting with 2003-07-10 and ending
in 2006-06-01, has zero ξ values except for the longest maturity (faiz456). A general economic
rule applying here is that the longer is the maturity, the more risk will emerge. Period 4 shows
zero ξ values for all maturity levels, which indicates that this period is associated with a lowest
interest rate risk period. This result appeared to be contradictory at the beginning because this
period is the first period of the impact of the global financial crisis on the Turkish economy.
An obvious question is, why does an economic crisis affect the risk associated with
interest rates in Turkey? An interpretation of this result can be found in the macroeconomic
implications of an economic crisis. During global crises, price levels and the level of interest
rates will generally decrease. Prices decrease because of lowering demand for goods, and the
rates of interest will also decrease for a similar reason. Compared to the earlier periods, in the
fourth period, the sensitivity of interest rates with respect to investment decreases, implying a
reduced demand for loans.
Figure 5 depicts the weekly observations of the total consumer loans and the claims under
legal proceedings between 2004-06-25 and 2010-01-01. It is very clear that the consumer
borrowing demand is slowing down for the latest period, and there is a dramatic increase in
the claims under legal proceedings. Hence the decrease in the risk of the rates of interest for
this period is smaller than expected.
These interpretations are statistically confirmed by the results of Table 3: Statistical
properties of interest rate changes for the four periods indicate that the results are significant.
Higher kurtosis means that a larger share of the variance is due to infrequent extreme deviation,
as opposed to frequent modestly sized deviations.
198
Bulletin of Monetary, Economics and Banking, January 2014
��
��
��
�
�
�
�
�
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
�
�
�
�
��
��
�
��
��
��
�
�
�
�
�
��
��
�
��
��
��
�
�
�
�
�
��
��
�
��
��
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
199
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
Figure 7. Total Consumer Borrowing And Claims
Under Legal Proceedings
Moreover, our results are also in line with the covered interest parity link. Table 3 depicts
the joint threshold exceedances for our four sub-periods in the Turkish economy. We have
used the dollar TL exchange rate as a proxy measure of the Turkish exchange rate. The joint
behaviour of changes in Turkish interest rates and the USD/TL is in line with our approach to
interest rate risk assessment to investigate the occurrence of joint daily threshold exceedances.
For each period, we define indicator variables as follows:
a USD-return exceedance happened on day t,
{ 10 IfOtherwise,
1 If an interest rate change exceedance happened on day t,
={
0 Otherwise,
Xt =
Yt
Here, we speak of a USD-return exceedance if the change in price of a USD in TL was
larger than its 90% quantile or lower than its 10% quantile, where the quantile is periodspecific. Likewise, an interest rate change exceedance is said to happen if the change in interest
rate is larger than its 90% quantile or lower than its 10% quantile, where quantiles are again
period-specific. Contingency tables for X and Y, together with their odds ratios, are shown in
Table 3. An odds ratio larger than 1 indicates a positive association of X and Y, that is, the main
diagonal of the contingency table has higher frequencies than expected under the hypothesis
that X and Y are independent.
Table 3 reveals that, as expected, a positive association was found for all periods, the odds
ratio indicating the strongest link for Period 1. Furthermore, a significantly positive association
(at a significance level of 5%) was only found for periods 1 and 3. Any interest rate exceedance
200
Bulletin of Monetary, Economics and Banking, January 2014
will also entail a change in interest rates. This link is similar for all four periods. Thus we observe
a covered interest parity condition in Turkish financial markets.
V. CONCLUSION
Measuring the interest rate risk is important for the emerging markets as well as the
globalised financial system. A risk hidden in the behavior of interest rates has not only directly
effect the functioning of markets but also have the power to alter the structure of the markets.
It is obvious that the normal distribution overestimates the probability of moderate changes and
underestimates the probability of large changes. This makes it inappropriate for risk analysis
in our case.
The use of the GPD in the context of absolute interest rates is well documented in
literature, our approach is different insofar as changes in interest rates constitute the target
of our analysis, reflecting the idea that risk arises from abrupt changes in interest rate rather
than in interest rate levels themselves. Our study clearly shows that the GPD, when applied
to interest rate changes, provides a good tool for interest rate risk assessment, and permit a
period-specific risk evaluation.
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
201
Appendix: Statistical Properties of Interest
Rate Changes
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
202
Bulletin of Monetary, Economics and Banking, January 2014
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Interest Rate Risk In Turkish Financial Markets Across Different Time Periods
203
REFERENCES
Andolfatto, David. 2012. Liquidity Shocks, Real Interset Rates, and Global Imbalances, Federal
Reserve Bank of St. Louis, Review, May/June 2012, 94 (3), pp. 187-95.
Ang, A., and Bekaert, G. 2002. International asset allocation with regime shifts. The Review of
Financial Studies 15, 1137–1187.
Bai, J., and Peron, P. 2003. Computation and analysis of multiple structural change models.
Journal of Applied Econometrics 18, 1–22.
Bali, T.G., and Neftçi, S.N. 2001. Estimating the term structure of interest rate volatility in
extreme values. Journal of Fixed Income, March 2001. Available at SSRN: http://ssrn.com/
abstract=262418.
Brigitte Desroches and Michael Francis. 2007. Global Savings, Investment, and World Real
Interest Rates, Bank of Canada Review, winter, 2006-2007.
Carneiro, L.A. F., and Sherris, M. 2008. Corporate interest rate risk management with derivatives
in Australia: Empirical results. Revista Contabilidade & Financas 19, 86–107.
Coles, S. 2001.An Introduction to Statistical Modeling of Extreme Values. Springer, Berlin.
Dai, Q., and Singleton, K.J, 2002. Speciﬁcation analysis of afﬁne term structure models. Journal
of Finance LV, 1943-1978.
Dufﬁe, D., and Kan R, 1996. Yield factor model of interest rates. Mathematical Finance 6,
379-406.
Fisher, R.A. and Tippett, L.H.C. 1928. Limiting forms of the frequency distribution of
the largest and smallest member of a sample,Proc. Camb. Phil. Soc.24, 80-190.
Gençay, R., and Selçuk, F. 2006. Overnight borrowing, interest rates and extreme value theory.
European Economic Review 50, 547–563.
Gençay, R., Selçuk, F., and Ulugülyağcı, A. 2002. High volatility, thick tails and extreme value
theory in value at risk estimation, Insurance: Mathematics and Economics 33, 337–356
Gençay, R., and Selçuk F. 2004. Extreme value theory and value at risk: Relative performance
in emerging markets. Journal of Forecasting 20, 287–303.
204
Bulletin of Monetary, Economics and Banking, January 2014
Hackbarth, D., and Miao, J., and Morellec, E. 2006. Capital structure, credit risk, and
macroeconomic conditions. Journal of Financial Economics 82, 519– 550.
Hosking, J. R. M., and Wallis, J. R. (1987), “Parameter and Quantile Estimation for
the GeneralizedPareto Distribution,” Technometrics, 29, 339-349.
Insel, A. 2003. The AKP and normalising democracy in Turkey. The South Atlantic Quarterly
102, 293–308.
Konstantin Borodin and Anton Strokov. 2011. Central bank Interest Rate and International
Trade in BRIC Countries: Agriculture vs. Machinery Industry?, conference paper on “Will
the „BRICs Decade continue?–Prospects for Trade and Growth”, Halle (Saale), Germany.
Meyfredi, J. 2005.Is There a Gain to Explicitly Modelling Extremes? Working Paper, Edhec Risk
and Asset Management Research Centre Publications, Nice.
Mirko Abbritti, Salvatore Dell’Erba, Antonio Moreno, andSergio Sola. 2013. Global Factors
in the Term Structure of Interest Rates, International Monetary Fund, working paper,
WP/13/223, November.
Paul Conway and Adrian Orr. 2002. The GIRM: A Global InterestRate Model , occasional paper,
Westpac Institutional Bank.
Philip Turner. 2014. The global long-term Interest Rate, Financial Risks and Policy Choices in
EMEs, BIS Working Papers, No. 441
Pickands, J. (1975), “Statistical Inference Using Extreme Order Statistics,” The
Annals of Statistics, 3,119-131.
R Core Team 2013R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.
Schmidbauer, H., and Rösch, A. 2004. Joint threshold exceedances of stock index returns in bull
and bear periods. Central European Journal of Operations Research 12, 197–209.
Stephenson, A.G. 2002. evd: Extreme Value Distributions. R News 2, 31-32.
Turanlı, M., andOzden, H.U., and Demirhan, D. 2002.SeçimTartışmalarının
Hisse Senedi Piyasalarına Etkisi, Istanbul Ticaret Universitesi Dergisi 2.
Woodford, M. 1999.Optimal Monetary Policy Inertia. NBER Working Papers 7261, Cambridge,
Massachusetts.
Zeileis, A., and Kleiber, C., and Kramer W., and Hornik, K. 2003. Testing and dating of structural
changes in practice. Computational Statistics and Data Analysis 44, 109–123.
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
205
The Impact of Capital Reversal and
the Threshold of Current Account
Deficit on Rupiah
M. Noor Nugroho
Ibrahim
Tri Winarno
Meily Ika Permata1
Abstract
This paper studies the effects of foreign capital flows toward the exchange rate of rupiah both in total
and across types of capital investment. This paper also analyzes the thresholds of current account deficit
which significantly affect the rate of Rupiah. The estimation shows the capital outflow affect the rate of
Rupiah to depreciate and is larger than the appreciation pressure of capital inflow (except when invested
in Certificate of Bank Indonesia, SBI). Furthermore, the rate of Rupiah is more sensitive on government
bond (SUN) than stock or SBI. The yield of this government bond largely affects the probability of the
capital reversal. Related to the current account, the estimation shows that after exceeds the threshold of
USD980 million monthly deficit orabout 2% of GDP, the exchange rate will depreciateby 12.7% (m-o-m)
with the lag effect of 4 months.
Keywords: Capital flows, exchange rate, current account deficit, threshold.
JEL Classification: F31, F32
1 Authors are researcher on Economic Research Bureau, Directorate of Economy Research and Monetary Policy, Bank Indonesia. The
views on this paper is solely of the auhtors and do not necessarily represent the views of Bank Indonesia. Corresponding author:
[email protected].
206
Bulletin of Monetary, Economics and Banking, January 2014
I. INTRODUCTION
The activity of external sectors in the economy is generally reflected on the balance of
payment dynamics, which later will affect the exchange rate of Rupiah. The main component of
the balance of payment is the current account (CA) and capital account (Capital and Financial
Account) and the change of foreign exchange reserve. When current account declines, it
commonly related to the reduction of competitiveness and export activity, or the increase of
import due to the strong domestic demands. After that, they then observe the other factors
such as the economic development of the trading partner, the import content in production
structure and in export products, etc. On the other hand, the decline of capital account is
sometimes linked to the falling of foreign capital’s value (sudden stop). In many countries, the
dynamics of the foreign capital flow is often associated with the flow of investment portfolio
considering the fact that this capital flow is more volatile than the other capital flows – direct
investment and foreign debt. Unlike current account of which activity is more determined by
fundamental and domestic factors, the dynamics of capital account depends more on investor’s
appetite (investor’s portfolio). Furthermore, the dynamics of portfolio investment can directly
affect the exchange rate2, therefore potentially cause instability in exchange rate and in domestic
stock market.
In its development, the studies which investigate the external balance are getting to be
more intensive following the Balance of Payment Crisis (BOP crisis), such as the crisis in Mexico
(1992), Asia (1997/8) and Russia (1998). The sudden stop and capital reversal phenomenon which
are linked to the deficit of CA (CA reversal) also occur in many countries. This phenomenon
is usually marked with the movement of the capital outflow in a gigantic amount, while the
capital inflow is limited. The imbalance of the capital outflow and inflow in one economy
directly influences their exchange rate. When the capital outflow is a lot bigger than the capital
inflow, the demands for foreign exchange will increase drastically and the exchange rate will
depreciate. On the next stage, the sudden stop can influence the current transaction activities.
Edwards (2004) observed 157 countries during the period of 1970-2001 and found 5.1% of
sudden stop events and 11.8% of CA reversals. Furthermore, he found from the total amount
of 2.228 observations, 46.1% experienced both sudden stop and capital account reversal;
showing that these two phenomena are related.
The sudden stop phenomenon which defined as the falling of the net inward flow of
foreign capital in a gigantic amount occurred in Indonesia in after the 1997/98 crisis. The large
of capital outflow increase the demand for foreign exchange much larger than the supply, which
put large pressure for Rupiah to depreciate. In some periods of sudden stop, Bank Indonesia
implemented a mix of policies which included foreign exchange intervention in considerable
amount, reflected on the decline of the foreign exchange reserve.
2 The influence of CA activity on exchange rate is indirect. This is because the foreign exchange resulting from export activity will
not directly enter the forex market. On the other hand, the needs of foreign exchange for import is not taken directly from foreign
exchange market as well. Instead of buying forex right at import transaction, the importer can buy the forex periodically when the
exchange rate is conducive for them.
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
207
After the crises 2008, the global capital flew to the developing countries with high
growth, supported by their strong domestic demand. However, the economic recovery process
of developed countries was full of uncertainties and made the finance market condition to be
susceptible to various negative sentiments. The finance market will fluctuate whenever the global
risk indicators are getting worse. For recipient developing countries, the worsening global risk
will make their foreign capital flow to be more volatile. The larger and the more volatile foreign
capital flow will cause instability in foreign exchange and financial asset prices.
Indonesia is one of few developing countries survived from the 2008 crisis and recorded
positive economic growth after the crisis. The foreign capital flow to Indonesia increased
significantly. The capital inflow reached the highest number of more than USD26.6 million in
2010 (USD13.2 million of it was portfolio investment). The capital inflow pushes the appreciation
of Rupiah for 12.5% to the average level of Rp9,080 per US dollar. However, the government
bond price (Surat Utang Negara, SUN) increased as its yield declined from 10.07% in the end
2009 to the lowest level of about 7.01 in October 2010 and closed at the level of 7.83 in
December 2010 (for 10 years SUN). The stock price index also increased from the level of 2,534
at the end of 2009 to the level of 3,703 by the end of 2010.
Indonesia in general enjoyed a net inflow condition. However, in several point in 2010
Indonesia also experienced capital reversal, for example in May where the net outflow was
significant and reached the number of USD4.8 billion, and USD865 million in November. The
sudden reversal commonly occurs when there is negative sentiment in the developed countries
who are still trying to get out of the crisis. The sudden reversal in Indonesia occur also on the
second half of the 2011with the net outflows of USD11.9 billion, and by mid of 2012 with
USD2.9 billion. Those events significantly affected the rate of Rupiah.
On the other hand, slow recovery of US and EU from the crisis and the falling of Japan
economy also influenced the decline of export activities in Indonesia. On the contrary, the
strong domestic demand still pushed the import to grow, resulting current account to decline
and even became deficit since quarter four 2011. Higher current account deficit affected the
exchange rate significantly, and when current account deficit is relatively low it does not affect
the exchange rate significantly.
Since the capital flow and current account activity can influence the movement of exchange
rate, it is important to do a more comprehensive research on it. Moreover, the movement of
the exchange rate will in turn affect the inflation which is the main task of Bank Indonesia.
On the other hand foreign capital flow may affect the financial management and the financial
system stability.
This paper measure the influence of foreign capital flow on the exchange rate, both in
total and across investment types; SUN, SPN, SBI and stock. Furthermore, this paper also measure
the influence of current account balance on the exchange rate and also test whether there is a
current account balance threshold which significantly affecting the exchange rate of Rupiah.
208
Bulletin of Monetary, Economics and Banking, January 2014
The next session of this paper outline the underlying theory and some previous studies.
Section three discusses the data, the model and the method. Section four explain descriptive
analysis, estimation result, and discussion, while section five outline the conclusion and close
the presentation of this paper.
II. THEORY
The standard theory of open economy commonly refers to Mundell-Flemming model which
combines the internal and the external balance. Internal balance is the equilibrium between IS
and LM curve, where IS is the expenditure curve (Y=C+I+G+NX) and LM is the money rill curve
(M/P). The external balance is shown by the balance of payment curve where the net position
of the current account and the capital account in total is zero.
The internal equilibrium is the combination of the real sector and the financial sector. IS
curve or the product market equilibrium is the accumulation of consumption, investment, and
real interest rates deduced by inflation expectation equation. On the other hand, investment
is the function of the income and the nominal interest rate. The net export is the function of
the exchange rate, domestic income, and trading partner income. The financial (factor) market
equilibrium is the function of income and real interest rate.
Consumption function is c = c[Y - T,r - E (p)]; investment function: I = I[i - E (p), Y-1]; net
export function: NX = NX (e,Y,Y*); and money market that represent the financial market is
M/P = f (Y,r); whereas the government expenditure is an autonomous factor (G=G).
�
�
��
��
��
�
��
�
��
�
��
��
�
Figure 1. Internal and External Equilibrium
The internal and external equilibrium is shown by point B on Figure 1 above which is
the intersection of IS curve, LM and BOP = 0 curve. The central bank policy to add the active
circulation will shift LM to LM’. Moreover, due to the capital inflow phenomenon, the capital
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
209
inflow to developing countries causes domestic currency demand to increase also. The result
is in point A, where the interest rate is lower and the acquired output is bigger. Some policies
which can be done are when the inflow is decentralised with the accumulation of reserve, it
causes point A to move back to point B. This will give implication to the longer capital inflow
period due to the higher interest rate. The second policy which is possible to be done is when
the central bank allows the entrance of capital and does not do sterilization so there is an
appresiation on exchange rate. This will cause the decline of competitiveness and export. It will
then shift IS from IS to IS’. Other policy which can be done is to do sterilization by limiting the
capital inflow. This causes the domestic efficiency to decline, so domestic investment requires
bigger cost (higher exchange rate) compared to the foreign one which offers lower cost.
In relation with external equilibrium, the exchange rate in floating exchange rate regime
is determined by the interaction between the supply and demand as the standard theory of
supply and demand. Demand is the amount of goods or service which can be bought by the
consumers in different price levels, where the higher the price, the fewer the amount of goods
or service the consumers want to buy. On the contrary, the supply is the amount of goods or
service which can be sold by the producers at various price levels. The spread between demand
and the amount of goods or service offered by the sellers is proportional with the price: the
higher the price, the more amounts of products offered by the sellers of the intended goods.
The equilibrium between the supply and the demand where the transaction take place is what
we call as a market.
The market is in the equilibrium condition when the interaction between the supply and
demand exists, and produces one equilibrium price on certain supply and demand quantities.
The foreign exchange market is another type of market where the traded commodity is foreign
exchange such as US dollar, Euro, or Japan Yen. The interaction between the foreign exchange
supply and demand which determine the foreign exchange rate is illustrated below.
��
��
��
��
�
��
�
��
��
��
��
��
��
��
��
��
��
��
��
Figure 2. Supply-Demand Foreign Exchange
210
Bulletin of Monetary, Economics and Banking, January 2014
The negative correlation between the demand for US dollars with its price in Rupiah (IDR/$)
is represented by ‘Demand for USD’ curve. The higher the exchange rate (the increase of US
dollar value), the lower the demand for US dollar will be. On the other hand, the correlation
between the exchange rate and US dollar supply is reflected in positive slope line. Figure 5
shows the equilibrium of USD at the rate of Rp9,000/USD. Due to certain effect – for example
when export increases, the supply of US dollar will shift to the right and create new equilibrium
at E2 (Rp8,500/USD). In this case Rupiah appreciate or strengthen relative to USD.
Since the monetary crisis in the 1990s, economists have focused their attention on the
global capital flow behavior. Some researches show that in globalization era, sudden capital
reversal or sudden stop gives negative impacts on domestic economy. According to some
economists (Dornbusch, Goldfajn and Valdes, 1995), sudden capital reversal also affects the
Current Account (CA) reversal and expensive economic adjustment. Most of economic crises
occurred in emerging market nowadays are marked by the sudden capital reversal (Calvo, 1998)
which is then followed by the significant decline of output.
Edwards (2004) did a research using the panel data to answer the phenomenon on the
impact of CA reversal toward the economic activity. They concluded that CA reversal will likely
take place if the CA deficit reaches 4% of GDP. The CA reversal is also influenced by other
factors such as foreign debt, domestic credit growth, and volatility of capital flow.
Sula (2008) used probit method and panel data to investigate the main factor causing
the sudden capital reversal. It indicated that the main cause of sudden capital reversal is the
high capital inflows in the previous period (around one to three years), and the composition of
capital inflows which is dominated by non FDI.
Kaminsky, Lizondo, and Reinhart (1998) built and estimated empirical model on currency
crisis by observing some variables with certain threshold and one of the variables are sudden
capital outflows. They found the level of exchange reserve of less than three months import
equivalent and a more than 5% GDP current account deficit will trigger the currency crisis.
Meanwhile, Frankel and Rose (1996) developed a method to provide early signal for exchange
rate crisis occurrence by using probit model for developing countries. They found that sudden
stop capital inflows and foreign debt composition significantly lead to currency crisis.
Sachs, Tornell and Valasco (1996) used panel data of 1995 to analyze the occurrence
of currency crisis which is known as ‘tequila effect’ after the Mexico crisis. The found the
significant effect of the vulnerable banking system, the overvalued exchange rate, and the
falling of national foreign exchange reserve.
Another research using independent variable with threshold was carried out by Kan and
Senhaji (2001). They used panel data and found supporting inflation threshold for the economic
growth was 1-3 percent for developed countries and 11-12 percent for developing ones.
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
211
Related to trade balance, Stahn (2006) carried out research on German exports and imports
both with European Union and outside European Union (particularly the United States) using
error correction model (ECM). He used REER and total sales deflator as price competitiveness,
and regional export as a proxy for the demand for German export. The result confirmed the
significance of price and demand on affecting the Germany’s export.
Chinn and Prasad (2003) analyzed determinant of current account in developed and
developing countries during the period of 1971-1995. Particularly for developing countries,
Chin and Prasad found that the surplus on government budget, the net foreign assets position,
and the trade volatility are closely related to current account balance.
Calderon et.al (2000) investigated current account deficit in 44 developing countries
during 1966 to 1995. The study found three conclusions; first, the higher the economic growth
of a developing country, the more likely that the occurrence of CA deficit to be higher; second,
the higher the economic growth of a developed country the more likely that the CA deficit
occurrence is lower; third, the higher the interest rate will result on the lower CA deficit happens;
and fourth, the appreciation of the real exchange rate increases the CA deficit, and (v) the
international interest rate which is slowly decreasing will lead to the rise of CA deficit.
On Indonesian case, Sahminan et.al (2009) carried out research on the determinant of
Indonesian current account as well as its dynamics using annual data from 1993 until 2008.
The consumption rate, the real exchange rate and investments are the factors which explain the
CA fluctuation in Indonesia. Moreover, by implementing inter-temporal approach, the research
also found that the CA fluctuation in Indonesia is optimum.
This paper attempts to contribute to the scientific literature which relates to the impacts
of sudden capital reversal phenomenon and CA on the exchange rate of Rupiahin the period
subsequent to the global financial crisis in 2008. The next session outline the data we use and
empirical model we estimate.
III. METHODOLOGY
3.1. Empirical Model
By the law of demand, an increase (decrease) of the commodity price will decrease
(increase) the demand for the commodity and on the other hand increase (decrease) the supply
of the commodity. The foreign exchange market also follows this law. The increase of US dollar’s
price toward Rupiah (depreciation of Rupiah) will reduce the demand of US dollar, and on the
other hand will elevate the supply; vice versa. The interaction of both can be formalized on the
following reduced form equations:
212
Bulletin of Monetary, Economics and Banking, January 2014
Foreign exchange supply:
ܳ௧ௌ ൌ ܽ଴ ൅ ܽଵ ܵ௧ ൅ ܽଶ ܺ௜௧ ൅ ‫ݑ‬௧
(1)
Foreign exchange demand:
ܳ௧஽ ൌ ܾ଴ െ ܾଵ ܵ௧ ൅ ܾଶ ܺ௜௧ ൅ ‫ݒ‬௧
(2)
D
where Qst and Q t are the foreign exchange supply and demand; St is the exchange
rate (standard quotation of the domestic currency per US dollar); and Xit is other explanatory
variables.
The market equilibrium exist when Qst = QDt at equilibrium price St. However, in forex
market bank determines the buying and selling exchange rate therefore supply and demand of
foreign exchange does not always in equilibrium (Qst = QDt). Those excess demand or supply will
be absorbed by the bank. Since the bank tends to be risk averse and is limited on its net open
position,then the excess within the market is minimized or limited. To suppress the excess, the
bank can change the selling/buying exchange rate offered to their customers. In this condition,
the market condition for the foreign exchange faced by the bank can be represented by the
following reduced from equations [(1)-(2)]:
ܳ௧ௌ െ ܳ௧஽ ൌ ܽ଴ െ ܾ଴ ൅ ሺܽଵ ൅ ܾଵ ሻܵ௧ ൅ ሺܽଶ െ ܾଶ ሻܺ௜ ൅ ‫ݑ‬௧ െ ‫ݒ‬௧
(3)
By isolating St we can get:
ሺ௕ ି௔బ ሻ
భ ା௕భ ሻ
ܵ௧ ൌ ሺ௔బ
െ ሺ௔
ଵ
భ ା௕భ ሻ
ሺܳ௧ௌ െ ܳ௧஽ ሻ ൅ ቀ
ሺ௔మ ି௕మ ሻ
ቁ ܺ௜
ሺ௔భ ା௕భ ሻ
൅ ሺ‫ݒ‬௧ െ ‫ݑ‬௧ ሻ
(4)
and after simplifying the coefficients we have:
ܵ௧ ൌ ܿ଴ െ ܿଵ ሺܳ௧ௌ െ ܳ௧஽ ሻ ൅ ܿଶ ܺ௜௧ ൅ ݁௧
(5)
The foreign exchange net supply (Qst - QDt) – in an open economy – might come from the
international trade transactions (export-import) and the capital flows among the countries. All
of those transactions are recorded on Current Account(CA) and Capital and Financial Account
(FA) in the balance of payment.
ܵ௧ ൌ ܿ଴ െ ܿଵ ‫ܣܥ‬௧ െ ܿଶ ‫ܣܨ‬௧ ൅ ܿଷ ܺ௜௧ ൅ ݁௧
(6)
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
213
This paper will focus on the study of the impact of capital reversal to the exchange rate
of Rupiah. This is the reason we use equation 6 – and FA is defined as the short term capital
flow (hot money)3. With a few modifications, we can highlight some important parts below:
• The asymmetrical influence of capital inflows and capital outflows: FA is separated into
outflows and inflows;
ܵ௧ ൌ ܿ଴ െ ܿଵ ‫ܰܫ̴ܣܨ‬௧ െ ܿଶ ‫̴ܷܱܶܣܨ‬௧ ൅ ܿଷ ܺ௜௧ ൅ ݁௧
(6A)
• Reversal influence based on the foreign divestment in SUN: FA is separated into foreign
investment/divestment in SUN and other capital flows;
ܵ௧ ൌ ܿ଴ െ ܿଵ ܷܵܰ௧ െ ܿଶ ሺ‫ܣܨ‬௧ െ ܷܵܰ௧ ሻ ൅ ܿଷ ܺ௜௧ ൅ ݁௧
(6B)
• Reversal influence based on the foreign divestment on the stock: FA will be divided into
foreign investment/divestment on the stock and other capital flows;
ܵ௧ ൌ ܿ଴ െ ܿଵ ܵ‫݇ܿ݋ݐ‬௧ െ ܿଶ ሺ‫ܣܨ‬௧ െ ܵ‫݇ܿ݋ݐ‬௧ ሻ ൅ ܿଷ ܺ௜௧ ൅ ݁௧
(6C)
• Reversal influence based on the foreign divestment on the stock: FA will be divided into
foreign investment/divestment on SBI and other capital flows:
ܵ௧ ൌ ܿ଴ െ ܿଵ ܵ‫ܫܤ‬௧ െ ܿଶ ሺ‫ܣܨ‬௧ െ ܵ‫ܫܤ‬௧ ሻ ൅ ܿସ ܺ௜௧ ൅ ݁௧
(6D)
• The reversal influence originated from foreign divestment on stock: FA will be divided into
foreign investment and divestment on SPN and other capital flows;
ܵ௧ ൌ ܿ଴ െ ܿଵ ܵܲܰ௧ െ ܿଶ ሺ‫ܣܨ‬௧ െ ܵܲܰ௧ ሻ ൅ ܿସ ܺ௜௧ ൅ ݁௧
(6E)
On the model specified above, the foreign capital flows variables invested in SUN, stock,
SBU and SPN will be estimated in three alternatives; first, the net flows or net (inflows minus
outflows); second, in the form of gross inflows and outflows (divided into 2 variables); and
three,in the form of gross outflows only. This method allows us to see specific influence of
each investment type/outlets on capital reversal occurrence.
Instead of observing the influence of sudden capital reversal towards the exchange rate,
equation 6 also observe the influence of CA activity towards exchange rate, particularly on the
efforts to determine the deficit of CA threshold which significantly trigger the occurrence of
exchange rate depreciation.
3 The short term capital flow includes foreign investment on SUN, stock, Certificate of Bank Indonesia or SBI, and also some portion
of foreign debt (assumed to be 50%).
214
Bulletin of Monetary, Economics and Banking, January 2014
To complete this research, we will estimate the probabilities of the sudden stop occurrence
which is influenced by the push (foreign) and pull factors (domestic).
ܴ݁‫݈ܽݏݎ݁ݒ‬௧ ൌ ߔሾܿ଴ ൅ ܿଵ ‫ܣܫܬܦ‬௧ ൅ ܿଶ ‫ܩܵܪܫ‬௧ ൅ ܿଷ ܷܵܰ௬௜௘௟ௗ ௧ ൅ ܿସ ܸ‫ܺܫ‬௧ ൅ ݁௧
(7)
Where Φ a is the normal cumulative distribution function; DJIA is the stock price index in
the US; IHSG is the composite stock price index in Indonesia; SUNyield is the return of investments
in SUN; and VIX is the volatility index which reflects the risk factor. DJIA and VIX represent the
push factors, whereas IHSG and SUNyield represent the pull factors.
3.2. Estimation Technique
This paper estimate the threshold of current account (CA) when it starts significantly
affects the rate of Rupiah against USD. To find the threshold, we use non-linear model approach
threshold autoregressive (TAR).
TAR is a regime-switching model which enables a variable to behave differently. Generally,
the amount of threshold is often unknown and must be estimated al together with the other
parameters. TAR model also accommodates the possibility that the length of the adjustment
process for the regime to change after certain period, d. The value of d is usually recognized
as delay parameter.
Enders (2004)4 and Chan (1993) give the guidance to gain super consistent threshold
value. Some of required conditions are:
1. Threshold should be placed within the range of observation period.
2. Estimate TAR model with different threshold levels and store the sum of squared residuals
(SSR), then choose the smallest SSR. Another criteria to choose the best model is using
smallest Akaike Information Criterion (AIC) and Schwarz Bayesian Criteria (SBC).
We apply the following threshold autoregressive (TAR) model:
௞
௞
ο݁௧ ൌ ൭ߙ଴ ൅ ෍ ߙଵ௜ ‫ܣܥܮ‬௧ି௜ ൱ ሺͳ െ ‫ܫ‬௧ିௗ ሻ ൅ ൭ߚ଴ ൅ ෍ ߚଵ௜ ‫ܣܥܮ‬௧ି௜ ൱ ሺ‫ܫ‬௧ିௗ ሻ
௜ୀ଴
௜ୀ଴
௟
௠
௡
൅ ෍ ߛ௜ ο‫ܦܴܫ‬௧ି௜ ൅ ෍ ߜ௜ οܴ‫ܭܵܫ‬௧ି௜ ൅ ෍ ߠ௜ ο݁௧ି௜ ൅ ‫ݑ‬௧
௜ୀ଴
௜ୀ଴
4 Walter Enders. 2004. Applied Econometric Time Series. Wiley.
௜ୀଵ
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
215
which represent:
௞
௟
௠
௡
௜ୀ଴
௜ୀ଴
௜ୀ଴
௜ୀଵ
௞
௟
௠
௡
௜ୀ଴
௜ୀ଴
௜ୀ଴
௜ୀଵ
൭ߙ଴ ൅ ෍ ߙଵ௜ ‫ܣܥܮ‬௧ି௜ ൅ ෍ ߛ௜ ο‫ܦܴܫ‬௧ି௜ ൅ ෍ ߜ௜ οܴ‫ܭܵܫ‬௧ି௜ ൅ ෍ ߠ௜ ο݁௧ି௜ ൅ ‫ݑ‬௧ ൱ ݆݅݇ܽ‫ܫ‬௧ିௗ ൏ ߛ
ο݁௧ ൌ
൭ߚ଴ ൅ ෍ ߚଵ௜ ‫ܣܥܮ‬௧ି௜ ൅ ෍ ߛ௜ ο‫ܦܴܫ‬௧ି௜ ൅ ෍ ߜ௜ οܴ‫ܭܵܫ‬௧ି௜ ൅ ෍ ߠ௜ ο݁௧ି௜ ൅ ‫ݑ‬௧ ൱ ݆݅݇ܽ‫ܫ‬௧ିௗ ൒ ߛ
The equation above is the application of TAR model on distributed lag model. On the above
model, the change of exchange rate (depreciation/appreciation) is function of current account
level (CA), change of interest rate differential, the global risk movement, and the adaptive
expectation on exchange rate. The time lag from the independent variable is represented by
i, where as the time lag from threshold variable is represented by d. Variable I is the dummy
variable, where It-d = 0 when CA level is smaller than the threshold γ, and It-d=1 when CA level
is equal or larger than the threshold γ.
We use monthly data from January 2008 until June 2012. The data includes the nominal
exchange rate in log change (De); current account level (LCA); interest rate differential change
(DIRD); and the global risk change (DRISK).
IV. RESULT AND ANALYSIS
4.1. Descriptive Analysis
The Dynamics of Balance of Payment
The Balance of Payment in Indonesia fluctuates dynamically. Prior to the 1998 crisis, its
current account was more often deficit and after the crisis it recorded surplus (Figure 3). The
dynamic represent the economic structure of this country. Any changes either structural in
nature or temporary will influence CA.
Structural change usually evolves gradually in a relatively long period and mostly relates
to the development of the domestic economy. On the other hand,the domestic changes
which are temporary in nature usually affect the balance of payment quickly in shorter period.
As small developing and open economy, for Indonesia, all changes in international markets
(commodity price and world trade volume) can give significant and immediate impact on its
domestic economy, including on its current account. Those external changes will be transmitted
via trading network.
216
Bulletin of Monetary, Economics and Banking, January 2014
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Figure 3.
The Development of Transaction (% GDP)
Figure 4.
The Development of the Main Indicator BOP
��
After 2000 or post the crisis, the management of macro policy was better. One of the
most fundamental changes was the floating exchange rate regime which can detect the signal
much faster when there is external side imbalance. During this regime however, the external
indicator were also suffers from several shocks as reflected on balance of payment (BOP).
On domestic side, one source of economic instability is the change in national political
system. Indonesian economy contains substantial politic uncertainties since the political
reformation in 1998. The crisis of national leadership which is also characterized by the changing
of president through the Special Session worries the foreign investors. The Standard and Poor
(S&P) rating reduced the investment rating in Indonesia in May 2001 and in November 2001.
The international geopolitics particularly the event of September 11, 2001 affected the domestic
economy and negative reaction for foreign investors within this country. The strings of political
and economic instability affected the economic recovery process until 2002.
The period where the external equilibrium in Indonesia run relatively normal without
significant political fluctuation and disturbances is the period after 2002. Statistic and BOP
data also showed improvement relative to before 2000. Consequently, the external equilibrium
fluctuation is mainly originated from external factor, hence relatively easy to compare. In general
there are three episodes of huge pressure on Indonesia balance of payment; they are in the
year of 2005, 2008, and in 2012.
We recognized the fluctuation in 2005 as mini crisis, where Rupiah depreciate drastically
to the level of Rp11,000 per USD. This is equivalent to a depreciation of 8,5% annually, which is
higher than in 2008. The main trigger of the fluctuation in 2005 was the increase of the world
commodity price, especially crude oil. The Fed’s responded by increasing his rate by 200bps
and made the foreign investors to relocate their investment portfolio to United States. With
the increase of the world crude oil price, the domestic fuel price also increased by more than
100% and triggered sharp rise of inflation.
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
217
In 2008, the source of fluctuation also came from external factor, particularly the raise
of commodity price and the subprime mortgage crisis in United States. In the last quarter of
2008, the Lehman Brothers collapse added the strings of crises in global financial market. The
massive foreign capital outflow put a high pressure on exchange rate and depreciate Rupiah
to Rp12,000 per USD.
The latest episode of external shock for Indonesian balance of payment was in 2012,
and was mostly caused by the crisis in Europe, and the decline of China and India’s economy.
The decrease of non-oil and natural gas commodities and the current rise of the world crude
oil price also played a role in lowering Indonesian trade as evident in first and second quarter
of 2012.
These three episodes of high external pressure on Indonesia balance of payment above
are well described on Figure 4. The accumulation of net capital and financial transactions also
decline along with the depreciation of Rupiah.
The Characteristics of Short Term Capital Flow in Indonesia
The component of balance of payment can be divided into two big components; the current
account balance (CA), and the capital and financial balance (KA). In general, the components
of CA contributing surplus are trade balance and transfer, while service and income contribute
deficit (see Figure 5). The roles of those four components are varied depending on the nominal
value. Since 2000-2011, the trade balance contributed most by 49%, while transfer transaction
only contributed 6%. Income and services contributed 25% and 19% consecutively.
On service trade balance, Indonesia as most of developing countries has weak
competitiveness within the service products. This is the reason for their services transaction to
��
��
��
��
��
��
��
�
��
��
�
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Figure 5.
The Development of CA Balance Component
��
Figure 6.
The Development of Transaction Portfolio and PMA
218
Bulletin of Monetary, Economics and Banking, January 2014
be deficit. The use of transportation and insurance services which is closely related to import
activities causes the decline of CA which is relatively faster when there is a leap of import.
On the other hand, income transaction is related to the profit of foreign company on
their investment; either in portfolio and direct investment (PMA). The raise of foreign interest
to invest in 2010 and 2011 was one of the reason for the increase of deficit income.
On capital and financial balance (KA), the foreign capital flow is relatively limited and
fluctuate. The accumulation of the capital flow was relatively stable from 2000 until 2008
(Figure 4). The significant rise occurred after 2009 along with the domestic economy recovery
(see Figure 6). However, the role of the portfolio capital is still dominant. Since the instrument
for foreign exchange transaction is still limited, and on the other hand the domestic foreign
exchange market is shallow, then the short term negative sentiment can easily trigger the
foreign capital flow, and then alter the Rupiah rate.
The global economic crisis in 2008 due to subprime mortgage problems in the US
leaded foreign capital flow to the emerging countries. The crisis which then spread to the
European Union and the world forced the developed countries to apply a quantitative easing
and expansionary fiscal policy to save their economy. The abundant global liquidity flew to the
emerging countries. However, due to the elevation of the global economic recovery uncertainties,
this capital flow was very volatile.
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
Figure 7.
CA and FA Balance
��
��
��
��
��
��
��
��
��
Figure 8.
Portfolio Investment
As we have witnessed, the foreign capital flow to Indonesia also tended to increase
continuously, but most of them are short term capital in the form of portfolio investment. After
the 2008 crisis, the capital flow raced along to Indonesia was recorded to amount of USD49.7
billion during third quarter 2009 until 2011 quarter three. USD29 billion or 58 percent of this
inflow was portfolio investment (PI). The incoming flow of this capital pushed up the Rupiah
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
219
financial asset price as reflected in the increase of capital stock index (Indeks Harga Saham
Gabungan, IHSG) and the decline yield of government bond (Surat Utang Negara, SUN).
Among several studies on the foreign capital flow to Indonesia after the 2008 global crisis,
Agung et.al (2011) confirmed the rise of foreign capital flow to Indonesia which is dominated
by the short term fund (PI). The implication is that the foreign capital flow in Indonesia tends to
be more volatile which is reflected by the rise of coefficient of variation in portfolio investment
(PI). Similar pattern was found during the global crisis of 2008. This research comprehensively
reveals the foreign capital flow to Indonesia both in the form of FDI, portfolio (FPI), as well as
Other Investment (foreign debt). Portfolio investment flows directly affect the demand and
supply of foreign exchange; hence the Rupiah rate.
Agung et.al (2011) further analyzed it through the investment outlet (SUN, stock and
SBI). The investment on SBI was the most volatile investment (temporary in nature or not
persistent) and became more volatile subsequent to the 2008 crisis. Bank Indonesia responded
by implementing minimum holding period on SBI for six months. The foreign capital flow in stock
market was also volatile. The volatility on this type of investment tended to increase after the
2008 crisis. Capital flow invested on foreign bond (SUN) is different. Unlike SBI and stock, the
foreign investments on SUN in fact became more persistent. It was parallel with the economic
activity in Indonesia which still runs well – including the stable fiscal deficit.
Furthermore, Agung et.al (2011) also showed that the foreign capital to Indonesia was
relatively short investment (1-2 months). Nugroho (2010) also shown that the foreign capital
duration invested in Indonesia is relatively short (one month). This is in accordance with the
fact that most of the foreign agent transacting in foreign exchange market are traders, and
not truly investor. They do the foreign exchange buying and selling in the short term period
to gain profit.
Another significant characteristic from the foreign capital flow is its asymmetrical impacts
on the exchange rate. Agung et.al (2011) proves that the capital outflow has bigger influence on
the exchange rate of Rupiah and has longer duration compared to the impacts of capital inflow.
Sugeng et.al (2009) also proves that the capital outflows influence (the rise of foreign exchange
demand or the decline of foreign exchange supply from the foreign agent) has bigger impacts
on the depreciation of Rupiah compared to the capital inflows impact on Rupiah appreciation.
Sugeng (2009) also proves that the impacts of foreign exchange transaction which is carried
out by the foreign agent has more significant impacts on the exchange rate of Rupiah than
the impacts of foreign exchange transaction by domestic agents.
This section has briefly described the actual condition of the market, the agent behavior
and the external condition of Indonesian economy. The next section will confront this reality
with the result of the estimated model.
220
Bulletin of Monetary, Economics and Banking, January 2014
4.2. Estimation Result
The empirical test result (Table 1) shows that the capital flow– both the inflows and the
outflows- significantly affects the rate of Rupiah. However, the impact is asymmetrical where
the capital outflows give bigger impacts than the capital inflows. The estimation shows that
whenever the capital flow out – especially capital reversal – in about USD100 million, the
Rupiah will depreciate by Rp3,2 on the same day. The Rupiah depreciation due to this capital
outflow will continue to the next day, though with smaller depreciation; see Figure 9 below.
The accumulative impact of USD100 million capital outflow will depreciate Rupiah by Rp16 in
about 15 days. On the other hand, the impact of same amount of capital inflow will appreciate
Rupiah by Rp3, with a three days lag.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
�
��
��
��
��
��
��
�
�
��
�
��
�
��
�
� ��
��
Figure 9.
Impulse Response of Exchange Rate towards Shock
Capital Reversal (USD100 million)
The effect of monetary operation by central bank using government bond (SUN) is in
accordance with the theory. A liquidity expansion when Bank Indonesia increases their buy
on SUN will depreciate Rupiah rate. However, this impact is not statistically significant. On
the other hand, the exchange rate level in the previous period largely determines the current
Rupiah rate.
The capital inflow in Indonesia in the form of portfolio investment is generally on SUN (the
long term government bond, up to 30 years), the SPN (short term government Treasury bill),
SBI and stock. A more intensive research shows that the effect of capital inflow differs across
those types of portfolio investment. Generally, the foreign capital flow – either in the form of
gross inflow, gross outflow, or net flows – invested in all investment types above (except stock),
significantly affect the rate of Rupiah. Capital inflows lead the Rupiahto appreciate, while capital
outflow lead Rupiahto depreciate, whereas net inflows lead Rupiah appreciation.
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
221
Based on our estimation (see Table 2), the foreign capital flow in SPN gives the largest
impacts on Rupiah’s rate. The impact of foreign capital flow on SUN is the second largest,
followed by stock, and finally SBI. Interestingly, the estimation result indicates the asymmetrical
influence between the inflow and the capital outflow. The estimated coefficient for the capital
outflow is larger than the inflow one, except for foreign investment in SBI. The next section will
describe the effect of capital flow on Rupiah rate for each types of investment.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
1. The Impact of Investment Flows on SUN toward Exchange Rate
Foreign capital flow placed in SUN significantly affect the movement of Rupiah rate,
either we measure it as capital inflow, outflow, or net capital flow. Its affect is proven to be
asymmetrical. A USD100 capital outflow will depreciate Rupiah by Rp13 based on the estimation
using the equation 2. A slightly different of Rp14 is estimated using equation 3. On the other
hand, the same amount of capital inflow will appreciate Rupiah rate by Rp7.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
222
Bulletin of Monetary, Economics and Banking, January 2014
When we use the net capital flow, we found smaller coefficient. An USD100 million net
capital inflow placed on SUN will appreciate Rupiah by Rp1,8 – Rp2,9. Accordingly, the same
amount of net capital outflow will lead Rupiah to depreciate by the same figures.
2. The Impact of Investment Flows on the Stock towards Exchange Rate
For stock investment, not all capital flow affects the Rupiah’s rate. The net capital flow
and capital outflow affected the exchange rate; however, the capital inflow does not. The
outflow of capital from the stock investment to the amount of USD100 million will depreciate
Rupiah by Rp11 – Rp12. Whereas the net capital inflow of the same amount will appreciate
Rupiah by Rp4.8.
When we use the by net capital flow (inflow minus outflow), the estimated coefficient
showed an USD100 million of net capital inflow will appreciate Rupiah by Rp3 – Rp5.3; and
vice versa, same amount of net capital outflow will depreciate Rupiah also by Rp3 – Rp5.3.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
3. The Impact of Investment Flows on SBI toward Exchange Rate
The foreign capital invested on SBI behaves slightly different with the other types of capital
flow investment. The capital inflow invested on SBI affect the exchange rate much stronger
than the capital outflow from SBI. A USD100 million capital inflow to SBI will appreciate Rupiah
by Rp3.6, while the net foreign investment inflows on SBI by the same amount will encourage
the appreciation of Rupiah by Rp2.9.
For this type of investment (SBI), we estimate three model variants. The result show the
net flow of Other Investment (OI) also affected the exchange rate. The net capital inflow (out)
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
223
to the amount of USD100 million will appreciate (depreciate) Rupiah by Rp3. The estimation
result above also shows that the buying/selling operation of SUN by BI significantly affects the
rate of Rupiah. If Bank Indonesia run monetary expansion by buying long term government
bond (SUN) to the amount of Rp1,000 billion, then Rupiah will depreciate by Rp18.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
4. The Impact of Investment Flows on SPN toward Exchange Rate
Similar to foreign capital invested on long term government bond (SUN), the foreign
capital investment on Treasury Bill (SPN) also significantly affected the exchange rate of Rupiah.
This is evident either when we evaluate the capital flow with inflow, outflow, or with net flow.
An USD100 million capital outflow will depreciate Rupiah by Rp24 – Rp25. Contrary to this,
the capital inflow with the same amount will only appreciate Rupiah by Rp21, showing an
asymmetric effect. When we use the net value, the result show an USD100 million net capital
inflow (out) on SPN will push the Rupiah to appreciate (depreciate) by Rp15.
The result of the estimation shows the impact of monetary expansion when Bank Indonesia
buy SUN significantly affect the rate of Rupiah. The buying of SUN of Rp1,000 billion will push
down Rupiah to depreciate by Rp19; vice versa.
224
Bulletin of Monetary, Economics and Banking, January 2014
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
5. The Impact of Current Account Activity on Exchange Rate
We estimate several variants threshold estimation using TAR model, from which we choose
the best threshold values on three models: AR-4, AR-5 and AR-6. The three models show the
existence of CA deficit threshold, and if the CA deficit level exceed the threshold, then the
effect on exchange rate becomes significant. On the contrary, when the level of CA deficit is
under the threshold, then its impact will likely be tolerable for the market without causing any
exchange rate fluctuation.
On model AR-4, the threshold of current account deficit is USD980 million, while in AR-5
model, the threshold is USD810 million. On the last model, AR-6, the current account deficit
threshold is USD730 million.
The result of the estimation in January 2008 until June 2012 period shows the occurrence
of CA threshold level to the amount of deficit of USD980 million (monthly) with 4 months lag.
The impulse response figure shows that whenever the shock level of CA exceeds the threshold
( > USD 980 million), the impact transmitted to the exchange rate with 4 months lag is 12.7%
(m.o.m). The cumulative effect within12 month year is the depreciation of Rupiah by 13% (figure
10 and 11). On the other hand, if the CA deficit does not exceed the threshold, the effect to the
change of exchange rate is relatively small. The impulse response show the impact of USD500
million CA deficit (monthly), and the cumulative effect within 1 year is only 0,4%.
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
��
225
��
��
��
��
��
��
��
��
�
��
��
��
��
�
�
��
��
�
��
��
Figure 10.
Impulse Response on the Decline of CA toward
Exchange Rate (Deficit USD980 million)
��
��
Figure 11.
Impulse Response Cumulative on the Decline of
CA toward Exchange Rate (Deficit USD980 million)
The second variant of the model AR-5 also cover the same period of January 2008 until
June 2012. The impulse response shows that when deficit of CA exceed the threshold of USD810
million (monthly), the impact will be transmitted to the exchange rate with 4 months lag is
depreciation of Rupiah by 32% (m.o.m). The accumulative impact after 1 year is the depreciation
of Rupiah by 29% (graph 12 and 13). If the CA deficit level does not exceed the threshold, the
effect of the change on exchange rate is also small. Cumulatively, the impulse response shows
the shock of USD500 million CA deficit will depreciate Rupiah by 0.3% within one year.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
�
��
��
��
��
Figure 12.
Impulse Response on the Decline of CA toward
Exchange Rate (Deficit USD810 million)
��
��
Figure 13.
Impulse Response Cumulative on the CA Decline
toward Exchange Rate (Deficit USD810 million)
226
Bulletin of Monetary, Economics and Banking, January 2014
On the last model, AR-6, the threshold of current account deficit is USD370 million. The
impulse response shows that a shock of USD370 million CA shock deficit will depreciate Rupiah
by 4.4% (m.o.m) with 4 months lag, and after one year, the cumulative effect is the depreciation
of Rupiah by 3.3% (figure 14 and 15). When the shock is less than the threshold-say USD500
million, the cumulative impact is only 0.2% after one year.
�
�
��
��
�
��
��
�
�
�
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
Figure 14.
Impulse Response to the Decline of CA toward
Exchange Rate (Deficit USD730 million)
��
��
Figure 15.
Impulse Response Cumulative on the Decline of
CA toward Exchange Rate (Deficit USD730 million)
6. The Determinant and Probabilities of the Sudden Stop Occurrence
To capture the capital reversal behavior, we apply probit model and this section outline
its estimation result. Probit model capture the probability on capital reversal occurrence by
observing the changes on some determinant factors such as the global stock price index (using
Dow Jones Industrial Average index or DJIA), the global risk factor (VIX index), the stock price
index in domestic market (IHSG), and the yield of the long term government bond (SUN).
We estimate three variants of econometrical model to explain the behavior of capital
reversal. The first equation uses DJIA and yield SUN as explanatory variables. The estimation
result shows the global stock price index negatively affect the probability of capital reversal
occurrence. This means a global stock price index decrease will increase the probability of capital
reversal.When the global stock price index decrease, investor would expect the domestic stock
price will be affected hence they will withdraw their investment from Indonesia.
The yield of government bond (SUN) also negatively related to the probability of capital
reversal; when the SUN yield decrease, the probability of capital reversal will increase. The yield
is negatively related to the bond price. The decline of bond yield is associated with the price
increase and leads the investor to gain profit by releasing their holding on SUN, particularly
when the bond price is already over value. This is in line with the recognized behavior of foreign
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
227
investors who attract more on short term profit. They buy the asset when its price is low and
sell it right away when its price increases. The government bond (SUN) by foreign investors is
identical with capital reversal.
On the second empirical model, we specify the yield of government bond (SUN) and the
risk index (VIX) as explanatory variables. The influence of yield SUN toward capital reversal is
relatively similar with the first model. On the other hand, the risk index also significantly affects
toward the tendency of capital reversal occurrence. A worsening risk condition reflected by the
increase of VIX index will raise the probability of capital reversal.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
The third model use the risk index (VIX) and domestic stock price index (Indeks Harga
Saham Gabungan, IHSG) as explanatory. Both VIX index and IHSG index significantly affect
the probability of capital reversal. An increase of risk index or stock price tends to increase the
probability of capital reversal.
V. CONCLUSION
The foreign capital flow particularly the short term significantly affected the rate of Rupiah.
The capital inflow will encourage the appreciation of Rupiah against USD and vice versa; however
the marginal effect is asymmetric. In general the capital outflow depreciate Rupiah’s rate more
than when same amount of capital inflow appreciate Rupiah. An exception is for capital flow
invested in Certificate of Bank Indonesia (SBI).
This paper analyze the behavior of capital flow across investment types, including long
term government bond (SUN), Treasury Bill (SPN), stock, SBI, and Other Investment. We found
that the exchange rate of Rupiah is relatively more sensitive toward foreign capital flow changes
228
Bulletin of Monetary, Economics and Banking, January 2014
invested on SUN than stock or SBI. This is quite reasonable since the investors in the stock market
can switch across stocks where the price movement provides them larger profit. Thus, the stock
release by the foreign investors does not always imply capital reversal. A different case is for
SUN, which price movement of each series tends to be linear, hence leaving the investors with
only 2 options either invest in SUN or reverse their capital out (capital reversal).
On the other hand, the investment on SBI is relatively more limited and is related more
to liquidity management. Moreover, SBI has a short span of maturity date (less than 1 year)
so the investors automatically have to release SBI when they reach the maturity date, and any
attempts to get it back in the primary and secondary market is constrained with limited supply.
The exchange rate is also sensitive toward foreign capital flow on SPN. However, considering
its limited volume, then its influence on the rate of Rupiah is also limited.
The foreign investors decision to invest in Indonesia is influenced by the level of investment
return and risk. This paper found that the probability of capital reversal is significantly affected
by global monetary assets (DJIA) and global risk factor (VIX), as well as domestic investment
return (IHSG and SUN yield). The SUN yield significantly affects the probability of capital reversal,
more than other explanatory variable, as far as the global risk is in good condition.
The fund flow originated from current account (CA) transaction is also significantly
influence the rate of Rupiah. Moreover, the impact of worsening CA activity on Rupiah’s rate
will be larger when it exceeds certain threshold level. The estimation result shows that whenever
the CA deficit exceed the deficit threshold of USD980 million (about 2% of GDP), the exchange
rate will depreciate by 12.7% (m-o-m) with a 4 months lag.
Following the above findings, we derive some recommendations to keep the stability of
Rupiah’s rate; first, improving the monitoring on foreign investments especially on SUN and
stocks. Second, increasing the monitoring of stock price index (IHSG), the SUN yield, and the
global risk indicator, and use them to predict the capital reversal. Third, increase the performance
of current account by any means to avoid its deficit exceeds the threshold. Fourth, to maintain
the stability of Rupiah’s rate when potential of capital reversal is detected, policy maker need
to maintain sufficient foreign reserve to enable them intervene on forex market.
The Impact of Capital Reversal and the Threshold of Current Account Deficit on Rupiah
229
REFERENCES
Agung, Juda, M. Noor Nugroho and Yanfitri (2011), “Arus Modal Jangka Pendek di Indonesia
Pasca Krisis Global: Karakteristik, Prospek dan Respon Kebijakan”, Bank Indonesia Working
Paper, Juni 2011.
Berg, Andrew, and Pattillo Catherine (1999)” Are Currency Crisiss Predictable? A Test” IMF
Economic Review, Jun 1999; 46,2.
Calderon, Cesar, Alberto Chong and Norman Loayza, 2000.” Determinant of Current Account
Deficit in Developing Countries,” The World Bank Policy Research Working Paper, No.
2398.
Chinn, Menzie D., and Eswar S. Prasad. 2003.” Medium-Term Determinants of Current Account
in Industrial and Developing Countries: An Empirical Exploration” Journal of International
Economics, Vol. 59, pp 47-76.
Calvo G (1998)” Capital Flows And Capital – Market Crises: The Simple Economics Of Sudden
Stops” Journal of Applied Economics 1:35-54.
Edwards, Sebastian (2004),” Financial Opennes, Sudden Stops and Current Account Reversals”
National Bureau of Economic Research, Working Paper 10277, January 2004.
Edwards, Sebastian (2005),” Capital Controls, Sudden Stops and Current Account Reversals”
National Bureau of Economic Research, Working Paper 11170, 2005.
Frankel, Jeffrey A., and Andrew K. Rose, (1996)” Currency Crashes In Emerging Markets: An
Empirical Treatment” Journal Of International Economics, Vol. 41 (nov), pp. 351-66.
Hansen, Bruce (1999), “Threshold Effects in Non-Dynamic Panels: Estimation, Testing, and
Inference,” Journal of Econometrics, Vol. 93, No. 2. pp. 345-68.
Hutchinson M, Noy I (2006),” Sudden Stops And The Mexican Wave: Currency Crisis, Capital
Flows Reversal And Output Loss In Emerging Markets” Journal Of Development Economics
79(1): 225-249, 2006.
Kaminsky, Graciela, Saul Lizondo, and Carmen Reinhart, 1998” Leading Indicators Of Currency
Crisis” Staff Papers, IMF, Vol. 45 (March) pp.1-48.
Khan, Mohsin S., and Senhaji Abdelhak S.2001” Threshold Effects In The Relationship Between
Inflation And Growth”, IMS Staff Papers, Vol. 48 No. 1.
230
Bulletin of Monetary, Economics and Banking, January 2014
Nugroho, M. Noor (2010), “Perilaku Aliran Dana Jangka Pendek Asing dan Pengaruhnya pada
Nilai Tukar Rupiah”, Bank Indonesia Working Paper, Juni 2010.
Radelet S, Sachs J (1998) The East Asian financial crisis: diagnosis, remedies, prospects” Brooking
Paper, economic Article 1:147-212.
Rodrik D, Velasco A (1999)” Short Term Capital Flows” National Bureau of Economic Research,
Working Paper 7364, 1999.
Rudi Dornbusch, Ilan Goldfajn, dan Rodrigo Valdes. 1995”Currency Crisis And Collaps”,
Brookings Papers on Economic Activity, pp. 219-70.
Sachs, Jeffrey, Aaron Tornell, and Andres Velasco, 1996” Financial Crisis In Emerging Markets:
The Lessons From 1995,” Brooking Papers On Economic Activity:1, Brooking Institution,
pp. 147-215.
Sahminan, Ibrahim, Yanfitri. 2009.“ Determinants and Sustainabality Indonesia’s Current
Account Balance”. Working Paper No. WP/09/2009. Bank Indonesia.
Stahn Kerstin, 2006” Has The Impact Of Key Determinants Of German Exports Changed? Results
From Estimations Of Germany’s Intra Euro-Area And Extra Euro-Area Exports” Deutsche
Bundesbank, Discussion Paper Series 1: Economic Studies No. 07/2006.
Sugeng, M. Noor Nugroho, Ibrahim and Yanfitri (2009), “Dampak Dinamika Penawaran dan
Permintaan Valuta Asing Terhadap Nilai Tukar Rupiah dan Perekonomian”, Bank Indonesia
Working Paper, Juni 2009.
Sula, Ozan (2010)” Surges and Sudden Stops of Capital Flows to Emerging Markets”, Open
Economic Rev (2010) 21:589-605.
Fund Management and The Liquidity of The Bank
231
Fund Management And the Liquidity
of The Bank1
Gantiah Wuryandani
Ramlan Ginting
Dudy Iskandar
Zulkarnain Sitompul 2
Abstract
This paper analyzes the liquidity of the banks, both precautionary and involuntary liquidity. We
apply dynamic panel estimation on individual bank data covering the period of January 2002 to November
2011. The result shows precautionary liquidity is more determined by the operation of the bank. On the
other hand, the involuntary liquidity is more affected by the financial system condition. Controlling the
size of the bank, the effect of the financial system condition and the macro economy is larger for the small
banks. Moreover, the monetary policy in the form minimum reserve requirement affects the precautionary
liquidity of the small banks; while the central bank rate is less influential to the bank liquidity.
Keywords: Banking, Liquidity, General Method of Moment
JEL classification: G21, G11, C33
1 Authors thank to VimalaDewi, AnggayastiHayu, Indri Tryana, and RizkaRosdiana for their great help on data and estimation. We
also thank to Dr. Telisa Aulia for their great discusiion.
2 Authors are researche on Center for Central Bank Education – Bank Indonesia; [email protected], [email protected], diskandar@
bi.go.id, [email protected].
232
Bulletin of Monetary, Economics and Banking, January 2014
I. INTRODUCTION
Bank is an intermediary institution that accepts from and channels funds to the society. The
performance individual banks and the banking system in aggregate will highly depend on how
the banks manage their assets and their liability.The banks manage their assets and their liability
to gain profit and to raise the company value within some rules. The rules include sufficient
liquidity, low risk and sufficient capitals. Therefore, the management of assets and liability are
highly related to the bank liquidity. According to Keynes (1936), the three motivations in holding
cash or liquidity are for transaction, precaution, and speculation. Furthermore, Edgeworth
(1888) with the square root of law of precautionary reserves principle, the liquidity reserve of
the bank will increase equivalently to the root of the number of transactions.
Liquidity is vulnerable and can be suddenly drained from a bank. The liquidity problems
at one bank can spread to other banks that will eventually pose a systemic risk. A shock can
create liquidity spiral that leads to the loss of liquidity and may form financial crisis. Learning
from the history, the banking crises lately was mainly due to the liquidity crisis that caused
the banks fail to pay their obligations. Within the framework of the financial system safety
net (FSSN), as also proposed by Bagehot (1873), the central bank as the lender of last resort
(LLR) provides temporary liquidity loans with specific requirements to maintain the stability of
banking system. The liquidity assistance is granted especially when the failure of bank may
cause contagion effect and lead a systemic risk. Goodhart (1987) states that there is no clear
distinction between the condition of illiquid and insolvent, but banks that need liquidity through
LLR are suspect to be in the process of becoming bankrupt.
In general, the liquidity reserve of a bank is a guarantee or a precaution over a possible
financial penalty due to the increasing withdrawals or an increase in the minimum reserve
requirement. Some banks choose a strategy to have excess liquidity to give a signal of strong
liquidity to the market. However, excess liquidity may also be interpreted that the banks
have a bad liquidity management and is sub optimal on managing their assets portfolio and
liabilities.
The excess liquidity can also be the result of poor infrastructure in the payment system and
inter bankmoney market. Di Giorgio (1999) argues that the level of financial system development
can be reflected by the participation cost within the financial system. In developed countries, the
cost to process information, project evaluation and monitoring of borrowers is relatively low.
This allows the banks to manage their liquidity with a relatively low liquidity reserves. On the
contrary, the country with poor payment systems and with limited infrastructure on interbank
money market tends to complicate the banks in managing their liquidity. This leads them to
hold higher liquidity reserve.
In micro-banks, the asset and liability management of banks concerns the aspects of
liquidity risk, market risk, trading risk, capital and fund raising, profit target and growth plans.
Generally the banks face three types of risk; credit risk (transaction, counterparty, concentration,
Fund Management and The Liquidity of The Bank
233
and settlement), market risk (interest rate, exchange rate, liquidity), and operational risks
(processes, systems infrastructure, human resources). The main focuses in asset and liability
management by the banks are liquidity risk, exchange rate, and interest rates. In this case,
the optimal liquidity is the liquidity that is able to create optimal revenue and prevent the
occurrence of liquidity risk. At macro level, from the central bank perspective, an optimal asset
and liability management by the banks is the one that create liquidity in accordance with the
target of monetary policy.
Currently, the management of assets and liabilities by the banks in Indonesia indicates a fair
amount of surplus liquidity3. This excess is absorbed by the central bank through monetary policy
with market or non-market approach. The market operations involve monetary transactions
between the central bank and the banks in order to lower or to increase the liquidity in the
market. This includes the selling or buying the government securities or Certificate of Bank
Indonesia.
In Indonesia, monetary policy tends to be contractionary to absorb the excess liquidity
in banking. This condition arises as the consequence of the bail out policy in financial crisis of
1998 when dealing with the bank run within banking system. The non-market approach to the
monetary policy is conducted through minimum reserve requirement which dictate the banks
to place minimum amount of their funds at central bank.
Ganley (2004) stated that the surplus of liquidity can cause problems for central banks on
transmitting his monetary policy. It can also create difficulties for the central bank to intervene
the currency market. Moreover, excess liquidity may disturb the balance sheet and the income
statement of the central bank. The distortion on monetary policy effectiveness will likely lead
to a problem on the financial sustainability of the central bank, particularly when the main
instruments for monetary policy is the central bank securities.
The main source of bank liquidity is capital inflows such as long-term foreign investment,
short-term portfolio investment, and the fiscal deficit financing. With free foreign exchange
regimes, the capital inflow to Indonesia either long-term or speculative, is a significant
determinant for national liquidity. Speculative capital inflow is more distortive for the financial
system and monetary stability. On the other hand, a long term capital inflow will support the
domestic economic growth and job opportunities.
This paper will analyze the impact of banks behavior in Indonesia in accepting deposit and
channeling funds on their liquidity. Furthermore, this paper will also identify the determinants
of banking system liquidity, and the role of monetary policy on the liquidity management of
the bank.
3 Surplus of liquidity occurs when the cash flow on the market exceeds the needs for reserves, (Ganley, 2004). Within market
equilibrium, this is an ex ante disequilibrium and tends to be persistent due to autonomous increase of liquidity on central bank.
234
Bulletin of Monetary, Economics and Banking, January 2014
The next section of this paper provides theoretical background and existing literatures
on liquidity. Section three present the data and the research methodology, while section four
discussed result and analysis. Section five present the conclusion and close this paper.
II. THEORY
2.1. Determinant of Bank Liquidity
In terms of micro-enterprises, the Bank for International Settlements (2008) defines liquidity
as the ability of the bank to finance the asset increase and meet liabilities without incurring
losses. Valla, Escorbiac and Tiesset (2006) and Vodova (2011) define liquidity as the ability
to meet cash liabilities, and can be distinguished into funding liquidity and market liquidity4.
Borio (1997, 2001) argues that it is necessary to distinguish between ex ante liquidity balance
before the central bank intervention and ex post after central bank intervention. Edlin and
Jaffee (2009) stated that high liquidity is due to the ‘credit crunch’ or the reluctance of banks
to channel the credit.
The development and liquidity conditions are not only affected by the bank’s business
activities but are also influenced by the money market. With a well functioning money market,
the banks can better manage their liquidity and avoid deficiency or excess liquidity. The
money market in a country is determined by the market structure, the available instrument,
development, regulation, and the market liquidity. The liquidity condition of the financial system
will determine the monetary policy taken by the central bank in order to achieve the target of
inflation and to maintain the sustainable growth momentum.
The interbank money market (PUAB) or also called interbank call money market is a place
for the banks to lendor to borrow funds to keep their liquidity. The transaction is short-term
and is used to deal with daily liquidity gap. The PUAB is executed over the counter (OTC) with
direct communication among banks through Reuter Dealing Monitoring System (RDMS). The
structure of interbank money market in Indonesia tends to be oligopoly and segmented, and
is very shallow. This condition makes the banks less flexible in obtaining and allocating optimal
liquidity. Most major banks tend to hold excess liquidity and more often serve as a lender. The
segmentation in the interbank market creates sub system of money market within them. In
this case, the same lender banks and same borrower banks will transact within their group
continuously. These conditions led to different levels of counterparty risk and variation across
segments including price disparity. Under conditions of tight liquidity, the segmentation of this
interbank market tends to be stronger with increasing counterparty risk. This condition tends to
encourage interbank rate to increase and the price disparity to widen. In this case, the motivation
of banks to not release liquidity grows higher in order to maintain adequate liquidity.
4 A funding liquidity (Valla, Escorbiac, and Tiesset, 2006) is the asset that is ready to be converted into cash to meet liabilities or for
operational activity, while market liquidity is defined as the activities of banks in trading assets shown by the ability of banks to sell
non-liquid assets.
Fund Management and The Liquidity of The Bank
235
The structure of the Indonesian financial markets has a very limited instrument of which
there is a short-term securities instrument, which is no more than one year, such as commercial
paper, certificates of Bank Indonesia, repurchase agreements, banker’s acceptances, and
certificates in the interbank money market instruments deposit. The ever shallow instrument in
PUAB encourages banks to manage short term liquidity by holding onto a limited variation of the
instrument. In general, banks tend to have instruments that are highly liquid with low risk like
government securities (sovereign), the central bank securities, and other short-term securities.
Besides market conditions, various regulations related to risk management and liquidity urges
banks to behave in certain way in managing both liquidity and asset and liability portfolios.
The money market is an outlet or the foremost means of most major banks in managing
liquidity. The liquidity condition of the bank will be directly reflected in money market both in
the volume of transactions and the dynamic of interest rate. The tight liquidity in the banking
is marked by the rising interbank rates and the widening spreads between the purchase and
the selling price. The tight level of bank liquidity is reflected in the loan rate (borrowing) and
financing rate (lending). The indicative rate for interbank market is reflected in JIBOR (Jakarta
Interbank Offered Rate), which is the average price of the quotations from the contributor banks.
These JIBOR were published through Reuters and Bloomberg and also reported by the bank on
daily basis on their daily reports (Laporan Harian Bank Umum, (LHBU). In addition to JIBOR, the
average price of all banks can be monitored via LHBU on the central bank website.
In Indonesia commercial banks participating in interbank money market possess wide
gap in assets and capital. By the end of 2011, the capital of the bank ranged from 0.15 - 54
trillion rupiahs, while their asset ranged from 0.17 trillion to 465 trillion Rupiahs. There are only
7 (seven) banks with asset above 100 trillion rupiah, while banks with assets below 1 trillion
rupiahs were 20 banks out of a total of 122 banks.
The results of Vodova’s research (2011) indicates that the liquidity of the banks in
Czech, measured by several indicators, were positively determined by capital adequacy ratio
(CAR), interest rate loans, the nonperforming loans (NPL), and inter-bank rate. On the other
hand, the financial crisis, inflation, and economic growth had a negative impact on liquidity.
The unemployment, interest rate margins, profitability, and the monetary interest rate do not
significantly affect the liquidity of the banks. Vodova measure the liquidity in his study with the
ratio of liquid assets (cash, demand deposits at the central bank) to total asset, bank liabilities,
and credit line with the other bank’s counterparties.
Shen, Chen, Kao, and Yeh (2009) conducted a study to determine the factors of liquidity
risk by using panel data from 12 countries. The result of the study indicates that liquidity risk
is affected by illiquid assets, external financing, supervision, regulation, and macroeconomics.
Liquidity risk is negatively correlated with the performance of the banks in countries with a
market-based financial system. On the other hand, in countries with bank-based financial
system, the liquidity risk is not associated with the bank’s performance.
236
Bulletin of Monetary, Economics and Banking, January 2014
Saxegaard (2006) states that banks will hold excess liquidity more than precautionary
reserve when the economy is in liquidity trap condition. Under these conditions the yield is too
low compared to credit intermediation cost, of which the returns from the funds placed at the
central bank is better than channeling credit. The result of a study by Aizenman Agenor, and
Hoffmaister (2000) indicates that the contraction in credit financing in Thailand after crisis,
caused by the phenomenon of supply resulting in involuntary excess liquidity.
Research by Bathaluddin et al (2012) stated that the tendency for the banks in Indonesia to
store excess liquidity is influenced by the need for currency fluctuations, economic growth, cost
of funds, and significant liquidity lag. The precautionary liquidity in the study was defined as the
ratio of the bank funds placed in central bank securities (excess liquidity) to the third-party funds.
On the other hand, involuntary liquidity is the residual of the estimated precautionary liquidity.
Pontes and Sol Murta (2012) found that excess liquidity occurs due to the weak development of
the financial sector where the interbank market is less efficient, a low diversification of financial
instruments, and weak credit intermediation due to the expensive costs.
The total liquidity of the banks in agregate will not change despite the change in the
liquidity ratio at bank level. However, these changes affect the composition of liquidity in the
presence of excess liquidity. Based on the research conducted by Keister and McAndrew (2009),
the amount of liquidity available in the bank is determined by central banks policy and does not
reflect their financing behavior. On the other hand, Ganley (2004) argues that some factors that
determine the liquidity of banks are beyond the control of the central banks. These include the
flow of their reserve to and from central bank,and the amount of money hold by the public.
Aspachs, Nier, and Tiesset (2005) conducted a study on bank liquidity in the UK using
quarterly data of individual banks from 1985 to 2003. The results showed that the greater the
support of central bank liquidity in the times of crisis, the lower the liquidity reserves held by the
bank. Most banks in the UK also tend to do a counter-cyclical liquidity strategy with low liquidity
back up when the economy grows. The liquid assets in this study consist of cash, reverse repo,
and commercial paper. The dependent variable is the liquidity ratio, which is measured by the
ratio of liquid assets to total assets, or the ratio of liquid assets to total deposits. The explanatory
variables consisted of Net Interest Margin (NIM), the profit, the credit growth, bank size, the
growth of Gross Domestic Product (GDP), and the short-term interest rates. Interest rates and
GDP have a strong influence on liquidity, as well as the opportunity for future financing.
On the other hand, Acharya and Merrouche (2010) analyzed the bank demand for
liquidity and settlement in UK as well as its effect on the interbank market before and after
the sub-prime crisis. The result showed that the bank in UK hold liquidity 30 percent higher
after the interbank suspension in August 9, 2007, showing precautionary action. This creates
a tight liquidity conditions and created crisis. The increased demand for liquidity by the banks
has raised the interest rate and potentially created systemic risk.
Fund Management and The Liquidity of The Bank
237
Some actions required to reduce the stress and the volatility of interest rate in interbank
money market are monitoring, early stress test, recapitalization for the troubled banks, and
increase the liquidity above the emergency liquidity reserves. The results of the study by Berger
and Bouwman (2009) shows that monetary policy does not have a significant effect on the
liquidity for large and medium-sized banks with 90 percent possession of the market liquidity,
or above. The monetary policy is instead effectively affecting the liquidity of small banks. In
addition, there is no significant difference between the effects of monetary policy during the
financial crisis and in normal times.
Henry, Birchwood, and Primus (2010) conducted a study to estimate the demand for
precautionary reserves and dynamic impact of involuntary reserve on monetary policy in Trinidad
and Tobago, by using the GMM (generalized method moments) and VAR (Vector Autoregression)
method. Their results showed that the bank holded excess reserve as precautionary action
against liquidity shortages. The spread between lending rates and policy rate negatively
affects precautionary reserves of the bank. In addition, the dynamics of involuntary reserves
are influenced mainly by fiscal operations. Similarly, a decreasing allocated credit during slow
economic growth tended to increase the liquidity of the bank.
Another study conducted by Pontes and Sol Murta (2010) using TSLS (two stage least
squares) show that the credit growth, the government securities and the crisis influence the
bank liquidity. High lending rate will disturb the bank intermediation, thus causing liquidity
accumulation.
2.2. The Role of Monetary Policy
The central bank’s monetary policy is implemented to maintain monetary stability in order
to control the national liquidity. By implementing liquidity management we expect to attain
sustainable economic development. In Indonesia, the central bank sets policy rate (BI rate) as
a reference for the market participants.
The instruments of monetary policy currently consist of Certificate of Bank Indonesia (SBI),
standing facility of Bank Indonesia, and the minimum reserve requirement (GiroWajib Minimum,
GWM). Bank Indonesia use SBI and term deposit for open market operation.
The maturity for Certificate of Bank Indonesia (SBI) was initially one month then in 2011
Bank Indonesia extended the holding period to six months. The maturity for term deposits is
longer up to nine months. On the other hand, the standing facility of Bank Indonesia set since
June 2008 is to fine tune the market operation particularly to control the overnight interest
rate (ON) within the interbank money market. This is helpful to keep the ON rate to move
around the BI rate (interest rate corridor), hence will ensure the effectiveness of monetary policy
transmission through the interest rate channel.
238
Bulletin of Monetary, Economics and Banking, January 2014
The standing facility consists of deposit and lending facilities with interest rates based
on BI rate plus or minus certain spread. When the bank holds excess liquidity, they can deposit
their excessive funds in Bank Indonesia using deposit facility; otherwise use lending facility when
they face a liquidity shortage. The interest rate corridor originally used symmetrical spread to
the BI rate movement. At the end of 2012 the spread between deposit facility and BI rate was
-175 basis points, while the spread between lending facilities amounted to +100 basis points
from the BI rate.
The non-market monetary policy instruments includes minimum reserve requirement
(GWM) rule the banks to place their funds in Bank Indonesia by certain percent of the funds
the collect from third parties. Nowdays, the reserve requirement policy is associated with the
Loan-to-Deposit Ratio (LDR). The aim is to optimize the use of liquidity by increasing the banking
intermediation to support the economic growth. In this case the banks are required to meet
the LDR of 78-100 percent, otherwise the central bank will impose additional minimum reserve
requirement. The central bank provides time deposit for the bank, which is limited up to 3
percent of their third-party funding. The incentive to deposit their fund is the BI rate minus 2.5
%, therefore will not burden the banks due to their loss of time value of money.
The management of assets and liabilities is very essential to keep the operation of the
bank. The liquidity conditions may affect the systemic risk and monetary policy transmission. The
uncertain withdrawal by depositors encourages the banks to choose best strategy on asset and
liability management to ensure they meet their laibilities. In addition, changes and volatility in
interest rates, and the exchange rate will determine the compliance with the conditions capable
of fund withdrawal liability, either suddenly or massively simultaneous.
Freixas, Martin and Skeie (2009) conducted a study of the efficiency in the interbank money
market fund allocation and the optimal policy of the central bank in the presence of liquidity
shock. The results of the research showed that distributional liquidity shock crisis will increase the
market segmentation across bank (disparity), and the central bank should lower the interbank
rate. The failure to lower interest rates in the times of crisis will worsen the financial stability
with the increasing probability of a bank run (simultaneous withdrawals by depositors).
On macro interest perspective, Saxegaard (2006) mentioned the need to distinguish
between precautionary excess liquidity with involuntary excess liquidity (excess liquidity that
exceeds the precationary). He found that in Sub-Saharan Africa countries (SSA), banks tend to
have excess liquidity that is involuntary due to the underdeveloped financial markets, the lack
of credit allocation, and the increasing government deposits at the bank. If a bank has excess
liquidity to meet the needs of anticipation (precautionary), the central bank does not need to
sterilize the economy since it potentiallytrigger inflation.
The behavior of the banks in SSA above may indicate a structural problem that causes
the inefficient allocation of funds. The involuntary liquidity in general serves as a secondary
reserve and is intended to address the possibility of liquidity gap in bank operation or the
Fund Management and The Liquidity of The Bank
239
likelihood of liquidity shocks. By holding the involuntary liquidity it naturally means the banks
pay opportunity cost to obtain income. In this case, Sacerdoti (2005) argues that SSA countries
need development of debtor information, accounting and auditing standards, and also legal
and regulatory framework. O’Connell (2005) also argues that involuntary excess liquidity will
disturb the mechanism of monetary policy transmission. Thus, an understanding of the sources
of excess liquidity becomes important to choose optimal monetary policy.
On his study about the pattern of excess liquidity banks in SSA countries above, Saxegaard
(2006) used SVAR (Structural Vector Autoregression). He also found the excess liquidity of the
banks weaken the mechanism of monetary policy transmission, and leave the monetary authority
unable to control the demand within the economy. Slightly in line with Saxegaard, Ganley (2004)
also stated that the liquidity absorption operation by the authority tended to use weak monetary
instruments such as central banks securities with high interest rates, making it less effective in
transmitting their monetary policy. In the long term, this raises important implications for the
central bank to finance the rising costs of monetary operations. This condition can seriously
affect the income of the central bank and its independency from the government. A continuous
loss of the central bank will require government recapitalization.
Meltzer (2009) and Feldstein (2009) argue that the increased liquidity can lead to an
inflationary pressure with rapid money creation through credit, and the central banks should
absorb this excess liquidity. However, Keister and McAndrews (2009) stated that the above
phenomena may occurs only when the central bank use a traditional monetary operation
framework. Currently, the Federal Reserve gives interest reward on bank liquidity deposited
at the central bank, thereby increasing market interest rates and restrain credit growth rate
without changing the amount of liquidity. By providing interest for reserves at the central bank,
the central bank may control the short-term interest rate which is independent from the level
of liquidity; therefore will not create inflationary pressures. In other words, the excess liquidity
in the bank does not always lead to inflationary pressures. A study by Bathaluddin, Adhi, and
Revelation (2012) using TVAR (threshold vector autoregression) indicates that there has been a
regime switching from the low liquidity to the high liquidity in Indonesia in 2005. Additionally,
excess liquidity caused ineffective monetary policy on controlling inflation.
Berger and Bouwman (2009) in his research show that monetary policy creates significant
effects of liquidity only to small banks. However, there is no significant difference of the impact
of monetary policy on liquidity creation during normal condition or crisis.
III. METHODOLOGY
The framework of this research is shown in the diagram below. The business activities
of the banks in collecting and allocating fund will affect their liquidity. On the other hand, the
liquidity conditions of the bank will affect the economic activity reflected on Gross Domestic
Product (GDP), hence the rate of inflation. The actual inflation and inflation expectations will
240
Bulletin of Monetary, Economics and Banking, January 2014
determine the monetary policy reaction of the central bank on controlling liquidity in order
to achieve his inflation target. The policy will influence macroeconomic conditions such as
interest rate movements, the exchange rates, and the economic growth. The changes of these
macroeconomic measures will be anticipated by the individual banks to choose strategy on
collecting and allocating fund. The sources of fund to the banks are demand deposits, savings
deposits, time deposits, loans, and capital. On the other hand, the allocation of fund may take the
form of cash, demand deposits at the central bank, demand deposits in other banks, securities,
loans, and other placements. All the process is inter-related and form continuous cycle.
Macroeconomy
Nilai tukar
Suku bunga
Krisis
Penempatan dana
Kas
Giro dibank sentral
Giro dibank lain
Surat berharga
Kredit
Penempatan lainnya
Penghimpunan dana
Giro
Tabungan
Deposito
Pinjaman
Modal
Transmisi
Moneter
Reaksi Kebijakan
Moneter
Likuiditas
Precautionary
Involuntary
PDB
Inflation
Diagram 1. Research Framework
We use monthly data of individual banks from January 2002 to November 2011 published
on financial report of banks. The samples include 110 banks from the total population of 122
banks. In this study, the sampling covers only conventional banks, considering that Islamic banks
have a different operational activities and different money market. We classify the banks based
on asset;large banks with asset above 50 trillion rupiahs, medium banks with asset of 10 to 50
trillion rupiahs, and small banks with asset lower than 10 trillion rupiahs.
Within our sample, the smallest bank had assets above 100 billion rupiahs, following the
provision of capital limit of minimum 100 billion rupiahs since 2010. The selection of number
Fund Management and The Liquidity of The Bank
241
of banks and period of observation meet the requirement that the cross section dimension is
larger than the period (N > T), which is useful on controlling individual heterogeneity when
there are unobservable behavior.
We specify a dynamic panel model and make use the estimation technique of generalized
method of moments (GMM). The selection of the best model is based on a panel data modeling
framework as shown in the diagram below.
Panel Data Model
Static Panel Data
HAUSMAN TEST
Fixed effects
Jika error
berkorelasi
dengan Xit
(regressors),
Random effects
Jika error tidak
berkorelasi
dengan Xit
(regressors),
Dynamic Panel Data
Jika lag dependent variable
digunakan: heterogeneity
Generalized Method of Moments
(GMM)
First difference
GMM
System
GMM
Diagram 2. Panel Data Model
The specification of empirical model for dynamic panel data model is below
Yit  Yi ,t 1   X it   it
(1)
Where Yit is the dependent variable (endogenous), Xit are explanatory variables (exogenous),
and mit are the residual. The instrument variable (IV) used is a certain lag of the endogenous and
exogenous of variables that are not correlated with the shock at time t. In general, symbols and
descriptions are used as variables in Table 1 appendix A.
Our endogenous variable is liquidity and we distinguish between precautionary and
involuntary as defined in Saxegaard (2006), Valla et al (2006) and Aspachs et al (2005). Yitis a
precautionary liquidity and Y2t is involuntary liquidity. The precautionary liquidity (Yit), is the ratio
of the cash plus minimum reserve requirements (RR), plus the placement on Bank Indonesia
and other banks (ODD), relative to total assets (TA). We specify the precautionary liquidity to
be more determined by the needs of the banks for operational activities.
Each bank needs different liquidity depending on their operation, market segments,
and tolerable risks. Based on focus group discussions, some banks consider the minimum
242
Bulletin of Monetary, Economics and Banking, January 2014
precautionary liquidity to be approximately 2-10 percent. The threshold for liquid to total assets
ratio for each bank depends on business operations, historical liquidity needs and risk appetite.
The involuntary liquidity Y2t is the ratio between central bank securities (CBSEC) plus government
securities (GSEC) and other securities (OSEC), relative to total asset (TA). The deposit of the
bank on central bank may take the form of Bank Indonesia certificates, term deposits and bank
Indonesia standing facilities. In general, the banks consider the threshold of involuntary liquid
asset ratios to be 15-18%.
Y1t 
Y2t 
C  RR  ODD
TA
(2)
i, t
CBSEC  GSEC  OSEC
TA
i, t
(3)
To explain the determinant of precautionary and involuntary liquidity, we use the balance
sheet component, representing the source and the allocation of the fund, monetary policy,
financial system and the macroeconomic condition. We focus more on variables that are
considered to affect the precautionary liquidity directly including their business activities. On the
other hand, the involuntary liquidity variable is driven more by market conditions and financial
systems, as well as the macro economy. We use the empirical specification below:
Y1i ,t  Y1i ,t 1  1RRRATE t   2 DTi ,t   3CREDITi ,t   4 FSI t   5ON t   i ,t
DTi ,t 
DD  S
 TD
i, t
i, t
i, t
TA
i, t
CREDITi ,t 
CRED
TA
i, t
(4)
(5)
(6)
i, t
wherei = 1,...., N is the observations and t = 1.... T is the monthly time dimension from
January 2002 to November 2011. Y1i,t-1 is the lag of the endogenous variable, RRRATE.t is the
level of reserve requirement, DTi,t in equation (5) is the ratio of deposits (including demand
deposits (DDi, t), savings (Si, t), and deposits (TDi, t) to total assets (TAi,t); CREDITi,t in equation (6)
is the ratio of credit (CREDi, t) to total assets (TAi, t). FSIt is financial stability index which represent
the pressure on financial system stability both in banking systems and capital markets; the ONt
is the interest rate for overnight interbank money market, and εi,t is error term.
Fund Management and The Liquidity of The Bank
243
The equation for involuntary liquidity (Y2t) is below.
Y2 i ,t   Y2 i ,t 1   1 BIRATE t   2 ER t   3 CREDIT i ,t   4 FSI t   5 CAR i ,t   6 GDPt   i ,t
DIT i ,t   4 FSI t   5 CAR i ,t   6 GDPt   i ,t
(7)
wherei = 1,...., N is the observations and t = 1.... T is the monthly period of January 2002
to November 2011. Y2i,t-1 is the lag of involuntary liquidity; BIRATEt is the interest policy rate;
ERt is the nominal exchange rate, CREDITi,t is the ratio of loans to total assets, FSIt is financial
stability index, CARit is the capital adequacy ratio or the percentage of minimum bank capital
requirement, GDPt is gross domestic product, and εi,t is the error term.
IV. RESULT AND ANALYSIS
4.1. Descriptive Statistics
We outline the statistics of variables used in this study in the Table 2. The high standard
deviation of involuntary liquidity shows that liquidity volatility involuntary (Y2t) tends to be
dynamic throughout the observation period range, as depicted on Figure 1. On the contrary,
the precautionary liquidity tends to have a relatively stable volatility compared to the involuntary
liquidity, as indicated with its low standard deviation.
The movement range of precautionary liquidity is relatively lower about 500 basis points.
This indicates that involuntary liquidity movements tend to follow the dynamic of the economic
situation which move from 1300 basis points (13 percent). At the time of 2005 mini-crisis
and the international financial crisis of 2008, the financial stability index (FSI) was above the
threshold 2, and involuntary liquidity decreased quite significantly. This indicates that with a
liquidity problem in the market, the involuntary liquidity will be used as primary buffer. Banks
tend to hold high reserve when the liquidiy of the financial system is relatively high, and then
use it when the financial system downturn and become more volatile. These behavior are
countercyclical to the market liquidity condition.
On the other hand, the volatility of the exchange rate (ERt) is relatively stable. A significant
surge was in late 2008 until mid of 2009 when the global financial crisis occurred due to the
subprime mortgage crisis in United States as illustrated in chart 3. The relatively stable exchange
rate movement is influenced also by the policy intervention by the Central Bank. Maintaining
the stability of the exchange rate is one of central bank target as mandated under the Central
Bank Act, in addition to the target on inflation.
Based on Figure 4, the policy rate does not affect the decision of the bank to allocate
their fund in central bank. When BI rate decline, the bank keep increasing their placement of
funds on central bank.
244
Bulletin of Monetary, Economics and Banking, January 2014
4.2. Precautionary Liquidity
The estimation result of precautionary liquidity using all banks is good and does not biased
upwards or downward (see Table 3 in Appendix D). The Sargan test statistics indicate that the
instrument variables (IV) used isvalid. The precautionary liquidity of all banks are significantly
influenced by previous precautionary liquidity, the minimum reserve requirement (RRRATEt), thirdparty funds (DTi,t), credit (CREDITi,t), financial stability index (FSIt), and interbank rates (ONt).
Across bank classification, our estimation indicates that the largest determinant of
precautionary liquidity are the level of previous precautionary liquidity, the third-party funds
(except for the medium-sized banks), the credit and the overnight rate of interbank market.
Banks still consider their historical liquidity conditions in determining the current one, and this is
in line with Bathaluddin et al (2012). The positive impact of third-party funds on precautionary
liquidity shows that when the third party fund increase, the bank face wider opportunities to
allocate funds including to expand their business. The level of allocated credit either in large
banks, medium, or small banks, negatively affect the level of precautionary liquidity; which is
consistent with Henry et al (2010) and Pontes and Sol Murta (2010).
Money market conditions reflected in overnight interbank rates (PUAB ON) negatively
influence the precautionary liquidity. The tight liquidity in the money market, as indicated by
the rising interbank rates will reduce the precautionary liquidity of the bank to overcome the
difficulty on obtaining liquidity. This is consistent with the results of Vodova (2011) and Acharya
and Merrouche (2010). On the other hand, the minimum reserve requirement (GWM) and the
financial stability index (FSI) do not significantly affect the precautionary liquidity, except on small
banks. This indicates that the non-market monetary policy using minimum reserve requirement
only affects the small banks, and is consistent with Berger and Bouwman (2009).
Small banks have limited activities hence will hold limited amount of liquidity reserve. This
is the reason why reserve requirement policy significantly affects their liquidity. On the other
hand, large and medium-sized banks hold higher liquidity. The amount is large enough to be
buffer for their operational acitivites; therefore a change in GWM will not affect their liquidity.
Furthermore, unlike the small banks, the large and medium-sized banks can easiliy fulfill their
liquidity needs from the interbank money market. In addition, when the banks deposit their funds
in central bank (maximum 3 percent of total third party funds), the central bank compensate
with 2 percent interest rate, and this is higher than the real cost.
Small bank liquidity is also affected by the stability of financial market. Estimation result
shows increasing FSI, which indicatesa lower stability of financial markets and usually followed
by a tight liquidity in the money market, will reduce the precautionary liquidity of the small
banks. These conditions show that the resilience of small bank liquidity is strongly depends on
the financial system condition, including the capital market.
Fund Management and The Liquidity of The Bank
245
4.3. Involuntary liquidity
The involuntary liquidity equation in Table 4 Appendix D shows that the involuntary liquidity
at all banks is determined by its own lag of involuntary liquidity (Y2i,t), the monetary policy rate or
BI rate (BIRATEt), the interbank rates (ONt), the exchange rate (ERt), credit (CREDITi,t), the capital
adequacy ratio (CARi.t), the financial stability index (FSIt), and the gross domestic product (GDPt).
The increase in interbank rates, exchange rates, and the FSI will reduce the involuntary liquidity
of the bank. The estimation on all sample showed that the increase in policy rate will increase
the involuntary liquidity of the banks. Generally the banks tend to use their involuntary liquidity
as a buffer to maintain liquidity in the event of financial market and financial system shock. It
is important to note that the estimation using medium-sized banks does not produce a good
equation despite not biased upward or downward. This has been indicated by the Sargan test
that the instrument variables used were not valid.
In general, involuntary liquidity largely depends on the previous involuntary liquidity. On
large bank sample, the previous liquidity is the only significant explanatory variable. The strong
influence of historical involuntary liquidity is in line with Bathaluddin et al (2012). The interbank
rates (PUAB) ON significantly affect the involuntary liquidity, except for the large banks. On the
other hand, the effect of capital adequacy ratio (CAR) is significant for small bank sample and
overall sample. The significance of capital requirement (CAR) on involuntary liquidity is also in
line with the study of Vodova (2011).
Using all samples, we find the policy rate has small effect on the bank liquidity. However,
this is not the case when we estimate the sample across their size (large, medium, and small
banks). The weak impact of the policy rate (BI rate) on involuntary liquidity hold by the bank is
in line with Vodova (2012). Policy rate is not the reason for the banks to deposit their fund on
central bank. The volume of the placement of the funds in the central bank securities continues
to increase, despite of the decrease of the policy rate. In addition, the flexibility of central bank
security as liquidity instrument is lower if the maturity of security is longer. The banks may face
high liquidity because of the speed of channeling credit is lower then the speed of third party
fund increase. On the other hand, the placement of funds in other financial product is still
limited due to several limitations to transact in foreign exchange and stock market, as well as
under-developed money market instruments.
The exchange rate only affects the liquidity of medium-sized banks, while the interbank
rates only affect the liquidity of the medium and small banks. The involuntary liquidity on small
banks affected by the lag itself, interbank rates, CAR, FSI, and it shows that the liquidity GDP.
The condition of involuntary small banks is also determined by the macroeconomic conditions
and macroeconomic finance. The condition systems such as FSI and GDP are only affecting
small bank liquidity, in line with the results of the study Aspachs et al (2005).
246
Bulletin of Monetary, Economics and Banking, January 2014
For the large bank sample, the previous involuntary liquidity significantly affects the current
involuntary liquidity. Other variables play minor impact on involuntary liquidity dynamics, which
indicates that major banks hold very high liquidity. Good liquidity resiliency of large banks
showed their better conditions on liquidity relative to smaller ones, and they tend to be lenders
in interbank money market. It is also supported by the fact that major banks are easier in raising
funds with lower cost than smaller banks. Furthermore, these large banks are equipped with
good infrastructure, networking, and more complete products, as well as better credibility.
Within this condition, smaller banks may necessary merger to better maintain their liquidity.
V. CONCLUSIONS
On this paper we devide the liquidity of banks into precautionary and involuntary
liquidity. Precautionary liquidity is the ratio between cash plus deposits on central banks and
other commercial banks, towards total assets; while involuntary liquidity is the ratio of tradable
securities (central bank securities, government securities, or other) towards total assets.
This paper showed that the accumulation and the management of the fund affect the bank
liquidity. Across the size of the bank, monetary policy and financial market condition (minimum
reserve requirement policy, interbank money market rate, and financial stability index) affect
more the precautionary liquidity of small banks. Furthermore, precautionary liquidity generally
depends on the operations of the bank, except for the small-sized banks.
The banks tend to use their involuntary liquidity as a buffer for their operational liquidity;
this is reflected by the dominant effect of financial market condition on the bank’s involuntary
liquidity. The monetary policy rate (BI rate) only affects the involuntary liquidty of the mediumsized banks, and not for the large and small ones. Furthermore, the macroeconomic conditions
such as financial stability index and gross domestic product, is also only affect the involuntary
liquidity of small banks.
Our conclusion above imply the central bank does not need to implement strict liquidity
absorption using policy interest rate, since the bank’s liquidity depend more on the operating
conditions, the capital, the financial system, and the macroeconomic condition.
Our result implicitely show that the banks tend to hold high liquidity involuntarily because
of the financial markets in Indonesia is shallow, and because of high uncertainty to get liquidity
from the market. Therefore, it is necessary to increase the depth of financial markets to expand
liquidity instrument. This will serve as good buffer for the banks and provide them flexibility on
managing their liquidity. The banks need to reduce their dependency on central banks, and
one way to do this is by reviewing the central bank standing facilities.
Fund Management and The Liquidity of The Bank
247
On the other hand, small banks have less immunity against the macroeconomic and
financial market condition. This implies a necessity for the small bank to merge and to combine
their asset, which will help them better manage their liquidity and increase their credibility. From
global perspective, this will also help Indonesian banking system to compete in international
markets such as forthcoming ASEAN Economic Community, 2015.
248
Bulletin of Monetary, Economics and Banking, January 2014
REFERENCES
Acharya, V.V, and Merrouche, O. 2010. Precautionary Hoarding of Liquidity and Inter-bank
Markets: Evidence from the Sub-prime Crisis. NBER working paper No.16395.
Agenor, P.R., Aizenman, J., and Hoffmaister, A. 2000. The credit crunch in east asia: What
can bank excess liquidity tells us? National Bureau of Economic Research, Working Paper
No.7951.
Alper, K., Hulagu, T., and Keles, G. 2012.An empirical study on liquidity and bank lending.
Central Bank of Turkey.
Aspachs,O., Nier, E., Tiesset, M. 2005. Liquidity, Banking Regulation and the Macroeconomy.
Evidence on bank liquidity holdings from a panel of UK-resident banks. Bank of England
Working Paper.
Bagehot, W.1873. Lombard Street, a Description of the Money Market. Henry S. King,
Bank for International Settlement. 2008. Principles for sound liquidity risk management and
supervision.
Bank for International Settlements. 2009. International framework for liquidity risk measurement,
standards and monitoring.
Bathaluddin, M.B, Adhi, N.M, dan Wahyu, A.W. 2012 Dampak Persistensi Ekses Likuiditas
terhadap Kebijakan Moneter. Buletin Ekonomi Moneter dan Perbankan Vol 14., No. 3,
Bank Indonesia.
Berger, A.N., dan Bouwman, C.H.2009. Bank liquidity creation, monetary policy, and financial
crises.University of South Carolina.
Borio, C. 1997. The implementation of monetary policy in industrial countries: A survey. Bank
for International Settlements, Economic Papers No.47.
Borio, C. 2001. A hundred ways to skin a cat: Comparing monetary policy operating procedures
in the United States, Japan, and Euro area. Bank for International Settlements, Monetary
and Economic Department Papers No.9.
Brunnermeier, M. K. 2009. Deciphering the Liquidity and Credit Crunch 2007-2008.Journal of
Economic Perspectives, Vol. 23, No. 1, pp. 77-100.
Di Giorgio, G.1999. Financial development and reserve requirements. Journal of Banking and
Finance.
Fund Management and The Liquidity of The Bank
249
Drehmann, M. and Nikolau, K. 2009. Funding Liquidity Risk. Definition and Measurement. ECB
Working Paper, No. 1024.
Edgeworth, F., 1888. The mathematical theory of banking, Journal of the Royal Statistical
Society, 51, 113-127.
Edlin, A.S. and Jaffee,D. 2009. Show Me the Money. University of California, Berkeley.
Freixas, X. and Rochet, J.C. 1998. Microeconomics of Banking. Massachusetts Institute of
Technology.
Freixas, X., Martin, A., dan Skeie, D. 2009. Bank Liquidity, Interbank Markets, and Monetary
Policy. Federal Reserve Bank of New York Staff Reports, no. 371.
Ganley, J. 2004. Surplus Liqudity: Implications for Central Banks. Lecture Series No. 3, Centre
for Central Banking Studies, Bank of England.
Goodhart, C. 1987. Why Do Banks Need a CB?. Oxford Economic Papers, p 75-89.
Henry, L., Birchwood, A., and Primus, K. 2010. The dynamics of involuntary commercial banks’
reserves in Trinidad and Tobago.XLII Annual Monetary Studies Conference.
Keister, T. and Mc Andrews, J. 2009. Why are banks holding so many excess reserves? Federal
Reserve Bank of New York, Staff Reports.
Keynes, J.M., 1936. The General Theory of Employment, Interest and Money.McMillan,
London.
O’Connell, S. 2005. Note on Excess Liquidity. Working paper, International Monetary Fund.
Pontes, G. and Sol Murta, F. 2012. The determinants of the bank’s excess liquidity and the credit
crisis: the case of Cape Verde. Faculdade de Economia da Universidade de Coimbra.
Sacerdoti, E. 2005. Access to Bank Credit in Sub-Saharan Africa: Key Issues and Reform
Strategies. Working paper, International Monetary Fund.
Saxegaard, M. 2006. Excess Liquidity and Effectiveness of Monetary Policy: Evidence from SubSaharan Africa. Working paper, International Monetary Fund.
Shen, C.H., Chen, Y.K., Kao,L.F., Yeh, C.Y. 2009. Bank Liquidity Risk and Performance.
Vodova, P. 2011. Liquidity of Czech Commercial Banks and Its Determinants.International
Journals of Mathematical Models and Methods in Applied Sciences, issues 6, volume 5.
250
Bulletin of Monetary, Economics and Banking, January 2014
Appendix A
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Fund��
Management and The Liquidity of The Bank
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
251
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
252
Bulletin of Monetary, Economics and Banking, January 2014
Appendix B
All data obtained from the statistic of the Bank Indonesia publication which are
longitudinal/monthly panel data from individual of conventional bank, monetary policy, money
market and foreign exchange and macro economy from January 2002 to November 2011.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Fund Management and The Liquidity of The Bank
253
Precautionary (Y1i,t) liquidity is the ratio between the summation of cash, account in BI and
account in other banks and total assets as described in equation (2). Involuntary (Y2i,t) liquidity
is the ratio between securities composite that is ready sold consisting of fund placement in
Central Bank in the shapes of securities, government securities and other securities and total
assets as described in equation (3). In this graph precautionary and involuntary liquidity are
liquidity aggregate.
��
��
��
��
��
��
��
��
�
� � ��
��
��
Figure 1. Precautionary (Y1i,t) and
Involuntary (Y2i,t) Liquidity
254
Bulletin of Monetary, Economics and Banking, January 2014
Appendix C
FSI (financial stability index) is composite index in money market consisting of banking
and capital market for measuring the durability of financial system. Involuntary (Y2i,t) liquidity
is the ratio between securities composite that is ready sold consisting of fund placement in
Central Bank in the shapes of securities, government securities and other securities and total
assets as described in equation (3).
�
��
��
��
��
��
�
��
��
��
�
��
��
�
��
� � ��
��
�
��
Figure 2. FSI and Involuntary Liquidity
BIRATEt is interest rate of monetary policy determined by Central Bank in order to an
open market operation. ERt is the nominal of exchange value.
��
��
��
��
��
��
��
��
��
��
��
��
��
�
�
��
�
��
�
�
� � ��
��
��
Figure 3. BI Rate and Exchange Nominal
�
Fund Management and The Liquidity of The Bank
255
CBSECi,t is bank fund placement in Bank Central such as in the shapes of term deposit,
securities of Bank Indonesia, and facilities of Bank Indonesia. BIRATEt is the interest rate of
monetary policy determined by central bank for the operation of open market.
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
�
��
�
��
�
�
� � ��
��
��
�
Figure 4. BI Rate and Bank Fund
Placement on Securities in Central Bank
�
256
Bulletin of Monetary, Economics and Banking, January 2014
Appendix D
This table shows the result of GMM (generalized method moment) on precautionary
liquidity (see table 1. for symbol and variable description). The model of precautionary liquidity
follows the equation (4). Instrument of variable (IV) which is used is dependent and independent
lag, in which the length of every IV of every model is different to get a significant result. The
number of samples for all banks are 110, big bank 17, medium bank 28, and small bank 65,
based on the amount of asset. Sampling for the period January 2002 until November 2011
uses the monthly data longitudinal panel.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Fund Management and The Liquidity of The Bank
257
This table shows the result of GMM (generalized method moment) to the determinant
(see table 1. For symbol and variable description) involuntary liquidity on equation (7). The
Definition of involuntary liquidity is in equation (3). Instrument of variable (IV) what is used is
dependent and independent lag, in which the length of IV every model is different to get a
significant result. Sampling for all banks are 110, big banks are 17, medium banks are 28, and
small banks are 65, based on the amount of assets. Sampling is for period from January 2002
to November 2011 by using the monthly data of longitudinal panel.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
258
Bulletin of Monetary, Economics and Banking, January 2014
Halaman ini sengaja dikosongkan
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
259
THE DYNAMICS OF TOTAL FACTOR PRODUCTIVITY
OFMEDIUM AND LARGE MANUFACTURING
IN INDONESIA
Ndari Surjaningsih
Bayu Panji Permono1
Abstract
This paper calculates and decomposes the Total Factor Productivity (TFP) for large and medium
scale industry in Indonesia covering the period of 2000-2009. By using Data Envelopment Analysis (DEA)
method, the result shows there is a shift of the supporting factors on the growth of TFP on manufacturing
sector within the 2 (two) sample period. In the period of 2000-2004, efficiency change becomes the
main contributor on the growth of TFP. Whereas in the period of 2005-2009, technical change becomes
the main supporting factor of TFP,however it goes along with the growth of negative efficiency change
or the decline of the company’s catching-up effect ability to adapt with the more advance technology.
The grouping of the sample across subsectors, technical change and also efficiency change shows the
declining amount of manufacture industry with superior productivity. Furthermore, the number of low
and weakening catching-up industry is increasing.
Keywords: Indonesian manufacturing, total factor productivity, technical change, efficiency change,
economic scale change, Data Envelopment Analysis
JEL Classification: L6, M11
1 Authors are researcher on Economic Research Group – DKM Bank Indonesia. The views on this paper are solely of the authors and
does not necessarily represent the views of Bank Indonesia; corresponding author: email here please.
260
Bulletin of Monetary, Economics and Banking, January 2014
I. INTRODUCTION
Sector of manufacturing in Indonesia is a strategic at least because of four reasons.
First, this sector is the largest contributor in Indonesian Gross Domestic Product (GDP). The
segment of this sector in GDP 2011 contributed 24.3 percent. Second, this sector absorbs high
employment, after the farming and the trading, hotel and restaurant, as well as service sectors.
Third, this sector is the main contributor in the total non-oil and gas export. About 38 percent
of the total export value or about 46 percent from the non-oil and gas total export in 2011
arise from manufacturing sector. Fourth, manufacturing sector has high backward lingkage and
forward linkage to other sectors. The linkage of this sector to other sectors, both the backward
and the forward linkage are above the average within all sectors.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
�
�
��
��
��
��
��
��
��
Figure 1.
Manufacturing Sector Segment in GDP
Figure 2.
Non-Oil and Gas Export Structure
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
The growth of manufacturing sector prior economic crisis of 1998 is relatively high 9.2%
(yoy) within the period of 1991-1998. However, the average growth declines after the 1998
crisis, which only reached the amount of 4.6% (yoy) in 2001-2011. Moreover, since 2004 the
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
261
growth slowed down but started to increase in 2010 and 2011. In general, the contribution
of manufacturing sector to economic growth declined in 2004-2009.
��
��
��
��
��
��
��
��
��
��
��
��
��
Figure 3.
Manufacturing Sector’s Growth
Considering the importance of manufacturing sector above, it is important to further
analysis the productivity of this sector, particularly the viability of its output. The use of Total
Factor Productivity (TFP) term on this paper includes the productivity of all production factors;
hence we do not analyze the productivity of individual factors as commonly found in many
literatures.
The first aims of this research is to calculate the Total Factor Productivity (TFP) of the large
and medium scale manufacturing companies in Indonesia; second, this paper will identify the
determinant of manufacturing productivity; and third, to analyze the technical change and
the efficiency change at subsector level. With these aims, we expect to be able to identify the
potential and the risk of manufacturing sector performance, as well as the policy required to
support it.
The second section of this paper will discuss on the theory, the third section will discuss
on the data and methodology, whereas part four will discuss on the result and analysis. The
conclusion will be provided in the fifth section and will close the paper presentation.
II. THEORY
2.1. The Concept of Productivity and Efficiency
The economic performance of a company can be reflected from the level of its efficiency
and productivity, that is the ratio of output towards input. The larger the output to input ratio,
the higher the performance of the firm will be. If the production process involves more than
262
Bulletin of Monetary, Economics and Banking, January 2014
one input, we need to aggregate the input with certain method and make index in order to
calculate the productivity ratio. The same case is required when the company produce multiple
outputs. This performance measure is a relative indicator across period or across competitors.
We need to clarify some terms related to productivity and efficiency; and the first one
is productivity. Productivity means the ratio of produced output towards the used input. This
productivity is reflected in the slope in a certain production point (case of one output y; and
one input x). As shown in figure 5. Company B has higher productivity than Company A.
�
��
�
�
�
��
�
�
Figure 4.
Illustration of Productivity
The second term is production frontier. The curve of production OF’ in Figure 5 shows
the amount of maximum output which can be produced in each input level. In other word the
production curve reflects the level of the use of technology by the company.
The third term is efficiency. This is a comparison of the output of certain company towards
the maximum output produced by other companies using same set of input. The company
is considered to be efficient when it operates exactly on the production line (frontier), which
�
��
�
��
�
�
��
��
�
�
Figure 5.
The Illustration of Efficiency
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
263
is at the point of B and C. On the contrary, it is considered to be inefficient if the company
operates under its frontier, which is at the point of A. At point A, the company still can improve
its efficiency to point B without additional input. In Figure 5. we can measure the A efficiency
of Company A as AA’/BA’.
We can use Figure 6 to distinguish efficiency and productivity. As explained above, the
productivity level is depicted by the slope of the straight line from point O. According to Figure
6, both Company A and B have equal productivity, however, the efficiency Company A is lower
than Company B. On the other hand, Company B and C have the same efficiency; however,
the productivity of Company C is larger than Company B. Thus, companies who have the
same productivity do not necessarily have the same efficiency, and the companies which equal
efficiency do not always have equal productivity.
��
�
�
��
�
�
�
�
Figure 6.
The illustration of Efficiency and Productivity Comparison
Allocative efficiency is another term. This is a combination of input composition which
produces output with minimum cost or yield maximum income. We can measure the allocative
efficiency only if we know the value of those inputs cost. We also know the term of technical
��
�
��
��
�
�
Figure 7.
Illustration of Technical Change
264
Bulletin of Monetary, Economics and Banking, January 2014
change, which can be estimated by how much the production frontier shift say from one period
to another. Figure 7 shows the technical change shifts from F0’ to F1’
Another important term is economies of scale. We can calculate this only if we release the
assumption of Constant Return to Scale; hence Variable Return to Scale. The value of economies
of scale is the distance between CRSand VRS. Figure 8 illustrates the economies of scale (grey
area) which is located between OB curve (Production Frontier under CRS) and OF’ (Production
Frontier under VRS). The optimal economic scale is the point where the company operates in
VRS frontier (OF’) with the highest productivity, compared to other companies within the same
OF’ curve. On this figure, the Company C is in its optimal scale.
��
�
��
�
�
�
�
�
Figure 8.
Illustration of The Economical Scale
The last term to clarify is Total Factor Productivity (TFP). This is a productivity which
includes all factors of production2 and can be decomposed into efficiency, technical change,
and economic scale. Thus, the concept of TFP used on this paper is different from common
method that measure TFP from residual (technology) in production function with capital and
labor as primary inputs.
2.2. Productivity and Efficiency Measurement
We have three options to measure efficiency; input oriented, output oriented, and distance
function. Using input oriented measure, we target certain output then minimize the use of
input. Within this method, the most essential variable to observe is input. On the other hand,
the output oriented method target certain level of input and then maximize the output.
2 Including calculating all of the output in the case of multiple-output production.
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
�
��
��
�
��
��
��
��
�
��
��
�
��
��
�
�
�
�
�
��
��
��
��
��
�
��
�
��
Figure 9. Input Oriented
265
�
��
��
��
��
��
Figure 10. Output Oriented
The third method, distance function, is commonly used on academic literature. Before
we explain this method in details, we firstly outline the production technology, which explain
the multiple-output production technology. The representation of technology set may refer to
Coelli (2005) following Fare and Primont (1995),
Let S be the technology set, while x and q represent Nx1 input vector and Mx1 output
vector. The vector value is non-negative real numbers in nature. Technology set below consists
of both input and output vectors (x,q) in which x produces q.
S = {(x,q): x can generateq}
Production technology can be represented with output and input set as follows.:
a) Output Sets, P(x), is an array of output vector, q, which can be produced using input vector,
x. Output set will be our base to construct production possibility curve (PPF) with two
output.
P(x) = {q: x can generate q} = {q : (x,q) ∈ S}
b) Input Sets, L(q), is an array of input vector, x, which generate certain output vector, q.
L(q) = {x : x can generate q} = {x : (x,q) ∈ S}
Without losing generality, we can explain multi-output technology with one input (x1)
and two outputs (q1 and q2). The input is function of these two outputs:
x1 = g(q1, q2)
The combination of two outputs produced by using certain level of input is our production
possibility curve (PPC). When the PPC curve is tangential to isorevenue curve, we have the
output combination which maximizes the revenue. The optimum point which produce maximum
266
Bulletin of Monetary, Economics and Banking, January 2014
revenue is point A, where the slope of isorevenue line (-p1/p2) is equal to the slope of PPC
curve.
��
��
��
��
��
�
�
��
Figure 11.
Production Possibility Curve and Maximum Income
In the case of multiple outputs, technical change may alter the production of certain
output relative to other output in two ways. From graphic below, we distinguish between
neutral dan non-neutral technical changes.
��
��
��
��
��
��
��
�
��
��
�
��
Figure 12. Technical Change Bias and Production Possibility Curve
We can apply the distance function both on output and input. The distance function
for a company operating in point A is the ratio of OA/OB (see Figure 13 and 14). Distance
function equal 1 (one) means the company already operates in PPC which correspond to certain
isoquant, L(q).
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
��
267
��
�
��
��
�
�
��
�
��
�
�
��
��
�
��
��
�
Figure 13. Output Distance Function
��
��
Figure 14. Input Distance Function
2.3. Malmquist Productivity Index
The term of productivity in this paper refers to the Total Factor Productivity (TFP) of
multifactors and not the partial factor productivity, such as labor productivity or capital
productivity. Partial measurement can be misleading when we measure the performance of a
company. When a company produces multiple outputs and multiple inputs, we can use the
profitability indicator, which is the ratio of total income against total cost from input.
For two companies, the TFP is measured by comparing the profit of the two companies.
After using the price of output and input, we can derive a simplified productivity measure as
equation below. Here we compare the productivity of the two companies using the real output
and the real input.
ߨ͸‫ݍ כ‬͸ Τ‫ݔ‬͸
ൌ
ߨͷ‫ݍ כ‬ͷ Τ‫ݔ‬ͷ
We can analyze the dynamics of productivity across period. Two periods comparison will
involve 2 (two) production technology sets, Ss and St; both for period s and t. Each technology
set relates to the vector of output qs and qt, as well as the vector of input xs and xt. Common
approach to do this is Malmquist Productivity Index (MPI), and we will apply this method on
this paper.
Malmquist Productivity Indexwas firstly introduced by Caves, Christensen and Diewert
(1982); a distance function method for representing technology in order to define families of
input, output and productivity indexes. For the output produced in period sand period t, there
is a technology that can produce maximum output using xs and xtinput. For example, if company
produce 80% of its maximum capacity using the input vector xs, andin period t he can produce
268
Bulletin of Monetary, Economics and Banking, January 2014
output 30% above maximum capacity using input vector xt, then the change of productivity
from period s to t is 1.30/0.80 = 1.625.
The calculation of MPI with using the technology reference in period s is:
݉଴ ௦ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻ
ൌ
݀‫ ݋ݏ‬ሺ‫ ݐݍ‬ǡ ‫ ݐݔ‬ሻ
݀‫ ݋ݏ‬ሺ‫ ݏݍ‬ǡ ‫ ݏݔ‬ሻ
If the company technically efficient (efficient) in both periods, then dos(qs, xs)= 1, thus:
݉଴ ௦ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻ ൌ ݀‫ ݋ݏ‬ሺ‫ ݐݍ‬ǡ ‫ ݐݔ‬ሻ
If we calculate MPI using the technology reference in period t, we use the following
formula:
݉଴ ௧ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻ
ൌ
݀‫ ݋ݐ‬ሺ‫ ݐݍ‬ǡ ‫ ݐݔ‬ሻ
݀‫ ݋ݐ‬ሺ‫ ݏݍ‬ǡ ‫ ݏݔ‬ሻ
With both periods MPI in hand, we can calculate the Malmquist TFP Index (MTFPI) as
geometrical average from both indexes as follows:
݉଴ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻ ൌ ሾ݉଴௦ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻ ൈ ݉଴௧ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻሿ଴Ǥହ MTFPI can be decomposed to 2 (two) components; efficiency change and technical
change. By using output orientated measure, the decomposition of MTFPI is:
݀௢௦ ሺ‫ݔ‬௧ ǡ ‫ݍ‬௧ ሻ ݀௢௧ ሺ‫ݔ‬௧ ǡ ‫ݍ‬௧ ሻ
቉
‫ ܫܲܨܶܯ‬ൌ ቈ ௧
ൈ
݀௢ ሺ‫ݔ‬௦ ǡ ‫ݍ‬௦ ሻ ݀௢௧ ሺ‫ݔ‬௦ ǡ ‫ݍ‬௦ ሻ
଴Ǥହ
In reality, a company often operates inefficiently, therefore dos(qs, xs) ≤ 1 and dot(qt, xt)≤
1. When the company is inefficient, the MTFPI can be calculated as:
݉଴ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻ ൌ
݀‫ ݋ݐ‬ሺ‫ ݐݔ‬ǡ ‫ ݐݍ‬ሻ ݀‫ ݋ݏ‬ሺ‫ ݐݔ‬ǡ ‫ ݐݍ‬ሻ
ቈ
݀‫ ݋ݏ‬ሺ‫ ݏݔ‬ǡ ‫ ݏݍ‬ሻ ݀‫ ݋ݐ‬ሺ‫ ݐݔ‬ǡ ‫ ݐݍ‬ሻ
‫ ݄݁݃݊ܽܥݕ݂݂ܿ݊݁݅ܿ݅ܧ‬ൌ ൈ
݀‫ ݋ݏ‬ሺ‫ ݏݔ‬ǡ ‫ ݏݍ‬ሻ
݀‫ ݋ݐ‬ሺ‫ ݏݔ‬ǡ ‫ ݏݍ‬ሻ
݀௢௧ ሺ‫ݔ‬௧ ǡ ‫ݍ‬௧ ሻ
݀௢௦ ሺ‫ݔ‬௦ ǡ ‫ݍ‬௦ ሻ
଴Ǥହ
݀௢௦ ሺ‫ݔ‬௧ ǡ ‫ݍ‬௧ ሻ ݀௢௦ ሺ‫ݔ‬௦ ǡ ‫ݍ‬௦ ሻ
቉
݄݈ܶ݁ܿ݊݅ܿܽ‫ ݄݁݃݊ܽܥ‬ൌ ቈ ௧
ൈ
݀௢ ሺ‫ݔ‬௧ ǡ ‫ݍ‬௧ ሻ ݀௢௧ ሺ‫ݔ‬௦ ǡ ‫ݍ‬௦ ሻ
ͲǤͷ
቉ The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
269
The
last MTFPI equation above can be decomposed into 2 components. The first component
evaluates the efficiency changes between period s and t, while the second component in brackets
estimates the changes of technology between the 2 periods.
Figure below illustrate productivity changes measures. It is assumed that a company has
a constant return to scale production characteristic with one input and one output. In period
s, the company operates in point D and move to E point in period t; both points are inefficient.
Efficiency and technical change correspondingly are:
‫ ݄݁݃݊ܽܥݕ݂݂ܿ݊݁݅ܿ݅ܧ‬ൌ ௤೟ Τ௤೎
݄݈ܶ݁ܿ݊݅ܿܽ‫݄݁݃݊ܽܥ‬
௤ೞ Τ௤ೌ
�
ൌቂ
௤೟ Τ௤್
௤ೞ Τ௤ೌ Ͷǡͻ
ൈ
ቃ ௤೟ Τ௤೎
௤ೞ Τ௤್
��
��
��
��
�
��
��
��
��
��
�
�
��
��
�
Figure 15. Malmquist Productivity Index
Considering that the calculation of MTFPI is based on CRS assumption, then there are
only two sources of productivity growth; efficiency change and technical change. However,
under variable returns to scale assumption, along with these two productivity sources, there
are also operating scale and efficiency scale. The weakness of MTFPI was covered later by
Grifell-Tatje and Lovell (1999) with generalised Malmquist Productivity Index by internalizing
the efficiency scale.
2.4. Data Envelopment Analysis
Data Envelopment Analysis (DEA) isa data oriented approach, and is used to evaluates
the performance of a set of entity called as DMU (Decision Making Units) which convert
multiple inputs to multiple outputs. The production frontier estimation with some numbers of
homogenous DMU uses non-parametric mathematical programming approach.
270
Bulletin of Monetary, Economics and Banking, January 2014
The first frontier estimation with piecewise-linear convex hull approach is pioneered by
Farrell (1957). Further development was carried out by Boles (1966) and Afriat (1972) using
mathematical programming method on frontier estimation. However, the DEA term become
popular after the work of Charnes, Cooper, and Rhodes (1978), who used an input oriented
model under Constant Return to Scale (CRS) assumption. After that, Banker, Charnes, and
Cooper (1984) modify this model using Variable Return to Scale (VRS) assumption.
Let there are N input and M output for each I companies. Each company is represented by
the vector of column xi and qi. The NxI input matrix X, and MxI output matrix Q, represents the
data for all of the companies. DEA model uses ratio; for each company, we need to calculate
the ratio of aggregate output toward aggregate input. In its aggregation, we use weight where
the optimum weight will be determined using mathematical programming. DEA model in the
form of Fractional Program (FP) is specified below:
݉ܽ‫ݔ‬௨ǡ௩ ሺ‫ݑ‬Ԣ‫ݍ‬௜ Ȁ‫ݒ‬Ԣ‫ݔ‬௜ ሻǡ
‫ݑݐݏ‬Ԣ‫ݍ‬௜ Ȁ‫ݒ‬Ԣ‫ݔ‬௝ ൑ ͳ
‫ݑ‬ǡ ‫ ݒ‬൒ Ͳǡ
݆ ൌ ͳǡʹǡ ǥ ǡ ‫ܫ‬
Optimum weight u and v in the FP above is obtained by maximizing efficiency subject
to the condition that efficiency value is less than or equal one. The problem arise from the FP
above is infinite solutions. Thus, the model in the form of FP above is converted into Linear
Programming (LP) as follows:
݉ܽ‫ݔ‬ఓǡ௩ ሺߤԢ‫ݍ‬௜ ሻǡ
‫ݒݐݏ‬Ԣ‫ݔ‬௜ ൌ Ͳǡ
ߤԢ‫ݍ‬௝ െ ‫ݒ‬Ԣ‫ݔ‬௝ ൑ Ͳǡ
ߤǡ ‫ ݒ‬൒ Ͳǡ
݆ ൌ ͳǡʹǡ ǥ ǡ ‫ܫ‬
The weight notation for FP and LP is distinguished to differentiate its mathematical
programming form. We can solve the formulation of DEA model in LP above, however, the
constraint will increase as the number of company is,and thuswe need to specify the LP form
above into Dual Programming (DP). In dual programming, the number of constraints will not
increase following the number of companies but rather only adding the variables to solve.
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
271
݉݅݊ఏǡఒ ߠǡ
‫ ݐݏ‬െ ‫ݍ‬௜ ൅ ܳߣ ൒ Ͳǡ
ߠ‫ݔ‬௜ െ ܺߣ ൒ Ͳǡ
ߣ ൒ Ͳǡ
We can specify the dual-programming using output orientated approach below (remaining
of this paper will use this):
݉ܽ‫ݔ‬థǡఒ ߶ǡ
‫ ݐݏ‬െ ߶‫ݍ‬௜ ൅ ܳߣ ൒ Ͳǡ
‫ݔ‬௜ െ ܺߣ ൒ Ͳǡ
ߣ ൒ Ͳǡ
On the DEA Constant Return to Scale (CRS), it assumes that all of the DMU operates
on the optimum economic scale. However, the existence of imperfect competition, financial
limitation, etc., make the DMU cannot operate on optimum economic scale. To deal with this
we can use DEA model under assumption of Variable Return to Scale (VRS). CRS model is not
starkly different with the VRS model except for the addition of convexity constraint (I1’λ = 1).
Below is the DEA model under VRS assumption:
݉ܽ‫ݔ‬థǡఒ ߶ǡ
‫ ݐݏ‬െ ߶‫ݍ‬௜ ൅ ܳߣ ൒ Ͳǡ
‫ݔ‬௜ െ ܺߣ ൒ Ͳǡ
‫ͳܫ‬ᇱ ߣ ൌ ͳ
ߣ ൒ Ͳǡ
The economic scale estimated from the above model does not indicate whether the
company is increasing or decreasing returns to scale. For this reason we impose a non-increasing
return to scale (NIRS) restriction to DEA model. If the technical efficiency in NIRS model is
different from technical efficiency in VRS model (TE NIRS ≠ TE VRS), then we may conclude it
as Increasing Return to Scale (IRS). On the other hand, if the TE NIRS isequal with TE VRS then
the case is Decreasing Return to Scale (DRS).
272
Bulletin of Monetary, Economics and Banking, January 2014
݉ܽ‫ݔ‬థǡఒ ߶ǡ
‫ ݐݏ‬െ ߶‫ݍ‬௜ ൅ ܳߣ ൒ Ͳǡ
‫ݔ‬௜ െ ܺߣ ൒ Ͳǡ
‫ͳܫ‬ᇱ ߣ ൑ ͳ
ߣ ൒ Ͳǡ
��
�
��
�
�
�
��
��
�
��
�
�
Figure 16. Ilustration on Non-Increasing
Returns to Scale
Estimation of dual programming model does not always provide the optimum efficiency
point. To ensure the solution of the model provide us optimum efficiency, we can use the
following model with slack variable:
݉ܽ‫ݔ‬ఒǡைௌǡூௌ െ ሺ‫ͳܯ‬ᇱ ܱܵ ൅ ܰͳᇱ ‫ܵܫ‬ሻǡ
‫ ݐݏ‬െ ߶‫ݍ‬௜ ൅ ܳߣ െ ܱܵ ൌ Ͳǡ
‫ݔ‬௜ െ ܺߣ െ ‫ ܵܫ‬ൌ Ͳǡ
ߣ ൒ Ͳǡ ܱܵ ൒ Ͳǡ ‫ ܵܫ‬൒ Ͳǡ
φ is parameter estimated from step one; OS is Mx1vector of output slacks; IS is Nx1vector
of input slacks; while M1 and N1 are column vector of ones with dimension of Mx1 and
Nx1consecutively.
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
��
273
�
�
��
�
�
��
��
�
�
��
Figure 17. Ilustration of Slack (input oriented)
In DEA, the estimation of Total Factor Productivity use index. Simple illustration, if a certain
company can produce the same output in period t and t+1, but in period t+1 only use 75% of
the input period t, then the TFP index will increase by 1/0.75=1.3. Similarly, if the company uses
the same input in both periods, but produces a 30% higher output at period t+1 compared to
period t, then the TFP index will be 1.3.
Beside the MTFPI explained in depth above, there are two other TFP index; Hicks-Moorsteen
TFP (HM TFP) Index, and TFP Index based on the Profitability Ratio. The earlier illustration use
HM TFP index with the following formula:
HMTFP Index =
Growth in Output
Growth in Intput
=
Output Quantity Index
Input Quantity Index
However, HM TFP index cannot explain the sources of the productivity growth (technical
change, efficiency change), and does not account for the price effect.
On the other hand, the second approach (Profitability Ratio) estimates the TFP index
using revenues and costs (price adjusted between period s and t). Similar with HM TFP index,
Profitability Ratio approach also neglect the price effects. For these reason, we will use Malmquist
TFP Index (MTFPI).
2.5. Research on the TFP of Manufacturing sector in Indonesia
There are some numbers of studies on productivity and efficiency in Indonesia. Generally
we can classify those researches based on their approach; Data Envelopment Analysis (DEA)
and Stochastic Frontier Approach (SFA). The first approach is non-parametric, while the latter
is parametric.
274
Bulletin of Monetary, Economics and Banking, January 2014
In the SFA group, Ikhsan (2007) examined the TFP growth and changes in technical
efficiency in Indonesian manufacturing industry during the period of 1988-2000. By using
Medium and Large Industry Statistic (SIBS), the study concluded that the average TFP growth
was 1.55 percent. The contributors of the TFP growth mainly came from technical progress of
about 1.89 percent, while the contribution of economic scale and technical efficiency are-0.13%
and -0.21% respectively. On technical efficiency changes, Ikhsan found a visible learning by
doing process in technology adoption because the company is not operating at its maximum
production capacity.
The National Development Planning Agency or Bappenas (2010) apply Ikhsan method
using 2000-2007 data also from SIBS. They found average productivity growth was approximately
0.22 percent. This productivity growth is lower relative to the productivity growth before the
crisis 19983. After slowed down during 2000-2004 probably due to post-crisis consolidation,
the growth of industrial productivity started to increase 2004-2007.
Bappenas found the major contributor of the productivity increase was technical efficiency
growth. On the other hand, the growth of technology and economies of scale contributed
negatively to the TFP, respectively -0.17% and -0.45%. In 2-digit level disaggregation in ISIC,
the Chemical sector recorded the highest TFP growth by averagely 0.21% per year, followed
by Non-Metallic Mineral sector (0.14%) and Food and Beverage sectors (0.09%). The lowest
productivity growth was in Wood Industry (-1.18%), Other manufacturing (-0.31%), and
Textiles sector (-0.08%).
Prabowo and Cabanda (2011) examined the productivity of Indonesian manufacturing
companies listed in Indonesia Stock Exchange 2000-2005 period. Using stochastic frontier
approach (SFA), Prabowo and Cabanda found technical inefficiency in the those companies. The
average technical eficiency was 0.7149, showing the company operated below its frontier.
Meanwhile, Saputra (2011) and Halim (2010) examined the productivity of the industrial
sector using DEA method. Saputra examine the level of technical efficiency of industrial
companies in Indonesia. By using the data in the UNIDO 3-digit ISIC level, they concluded
that there are 5 sub-sectors with highest efficiency, Tobacco; Iron and Steel; Transportation
Equipment; Non-Ferrous Metal; and Chemistry. In general, the efficiency of basic industry was
higher than the traditional industry in the category of low and high-tech industry. Nevertheless,
the latter industry showed higher efficiency in the last 2 years of their observation.
Halim (2010) specifically examined the marketing productivity and the profitability of
the company in Indonesia. The inputs used in this study were limited to equity and marketing
expenses. DEA method was applied to the five categories of industrial manufacturing company
listed on the Stock Exchange during 2001 to 2007; Food and Beverage; Clothing and Textile
Product; Plastics and Articles of glassware; Automotive Products; and Pharmaceuticals. The
3 Period1988-1992 and 1993-1996.
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
275
main conclusion of this study is the highest productivity of marketing was in 2005-2006, and
the major contributor was technological efficiency. A total of 44 companies were identified
operating at their efficient level. Based on their categories, Automotive had the highest
productivity and technical efficiency. The TFP for efficient company was positively related to
Return on Assets (ROA), reflecting the higher efficiency of marketing productivity, the higher
financial performance will be.
III. METHODOLOGY
3.1. Methodology
On this paper, the estimation of TFP growth and its components is based on the Malmquist
Indexand the application of DEA-Dual Programming method. The components of Total Factor
Productivity include technological changes, change of efficiency and change of economic of
scale, across companies, sub-sector, and across year.
Output Oriented Malmquist DEA model can be defined as follows:
݉ܽ‫ݔ‬థǡఒ ߶ǡ
‫ ݐݏ‬െ ߶‫ݍ‬௜ ൅ ܳߣ ൒ Ͳǡ
‫ݔ‬௜ െ ܺߣ ൒ Ͳǡ
ߣ ൒ Ͳǡ
Where ϕ is a proportional increase in output produced by company i, given constant
input; λ is the weight for each input or output; q is the output of company i; and Q is the
output of remaining companies. On the other hand, x is the input company i, while X is the
input for remaining companies.
The Malmquist TFP index is defined as follows:
݉଴ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻ ൌ ሾ݉଴௦ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻ ൈ ݉଴௧ ሺ‫ݍ‬௦ ǡ ‫ݍ‬௧ ǡ ‫ݔ‬௦ ǡ ‫ݔ‬௧ ሻሿ଴Ǥହ
Furthermore, the components of Total Factor Productivity Malmquist are derived from
the breakdown of Malmquist index, as follows:
ܶ‫ ݄݁݃݊ܽܥݕ݂݂ܿ݊݁݅ܿ݅ܧ݈ܽݐ݋‬ൌ ݀௢௧ ሺ‫ݔ‬௧ ǡ ‫ݍ‬௧ ሻ
݀௢௦ ሺ‫ݔ‬௦ ǡ ‫ݍ‬௦ ሻ
଴Ǥହ
݀௢௦ ሺ‫ݔ‬௧ ǡ ‫ݍ‬௧ ሻ ݀௢௦ ሺ‫ݔ‬௦ ǡ ‫ݍ‬௦ ሻ
቉
݄݈ܶ݁ܿ݊݅ܿܽ‫ ݄݁݃݊ܽܥ‬ൌ ቈ ௧
ൈ
݀௢ ሺ‫ݔ‬௧ ǡ ‫ݍ‬௧ ሻ ݀௢௧ ሺ‫ݔ‬௦ ǡ ‫ݍ‬௦ ሻ
276
Bulletin of Monetary, Economics and Banking, January 2014
‫ ݄݁݃݊ܽܥݕ݂݂ܿ݊݁݅ܿ݅ܧ‬ൌ ௧
݀௢௩
ሺ‫ݍ‬௧ ǡ ‫ݔ‬௧ ሻ
௦
݀௢௩ ሺ‫ݍ‬௦ ǡ ‫ݔ‬௦ ሻ
଴Ǥହ
௧
௧
௦
௦
ሺ‫ݍ‬௧ ǡ ‫ݔ‬௧ ሻȀ݀௢௖
ሺ‫ݍ‬௧ ǡ ‫ݔ‬௧ ሻ ݀௢௩
ሺ‫ݍ‬௧ ǡ ‫ݔ‬௧ ሻȀ݀௢௖
ሺ‫ݍ‬௧ ǡ ‫ݔ‬௧ ሻ
݀௢௩
ൈ
቉
‫ ݄݁݃݊ܽܥ݈݁ܽܿܵܿ݅݉݋݊݋ܿܧ‬ൌ ቈ ௧
௦
௦
௧
݀௢௩ ሺ‫ݍ‬௦ ǡ ‫ݔ‬௦ ሻȀ݀௢௖ ሺ‫ݍ‬௦ ǡ ‫ݔ‬௦ ሻ ݀௢௩ ሺ‫ݍ‬௦ ǡ ‫ݔ‬௦ ሻȀ݀௢௖ ሺ‫ݍ‬௦ ǡ ‫ݔ‬௦ ሻ
3.2. Data, Variable, and Proxy
The main data used in this study is the Survey of Large and Medium (Survei Industri Besar
dan Sedang, SIBS) published by the Central Bureau of Statistics (BPS). The period covers 20002009. Each company (KIPN) is grouped based on 3-digit ISIC code. For each company, we use
the following set of variables: output, capital, labor, raw materials, and energy.
We can calculate the output of the firm based on their production or sales. This study
use the first proxy, since the firm use all resources (capital, labor, raw materials and energy) to
produce a number of outputs despite of being sold or stored as inventory. The production value
will be deflated using Indonesian wholesale price index correspondingly for each sub sector.
The proxy for capital is estimated fixed or durable asset including land, buildings,
machinery, vehicles, and other durable asset. Some missing capital data during the survey
is estimated. We use capital at year t to estimate the capital for other years, using the firm
investment (purchase or maintenance), the value of sales, and depreciation (assuming equal
to 14%) during those missing period. To get real capital data, we use Gross Domestic Fixed
Capital Formation (GFCF) deflator to deflate the nominal capital.
For employment data, we consider the use of working hours to be appropriate, since same
amount of labor in a firm may generate different output when their working hours change
(due to overtime or temporary production stops). However, due to a lack of data, this study
used the number of workers as proxy for employment.
The intermediate input included raw materials and supporting materials originated both
from domestic and import. This intermediate input is deflated using wholesale price index,
which is assumed to be equal across companies.
For energy data, we use of fuel and lubricants as well as electricity. Both of these energy
were deflated before aggregating them to one energy composite. We deflate fuel and lubricants
using their corresponding wholesale price index (premium, kerosene, diesel fuel, diesel oil, fuel
oil, and lubricants). For electricity we use sectoral GDP for electricity as deflator.
Our data covers 49 subsectors with the total of 3,295 firms. The summary of variable
and the distribution of sample is provided on Table 2 and Table 3.
277
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
� ��
��
� ��
��
��
��
��
��
��
��
��
��
� ��
� ��
��
��
��
� ��
��
��
��
� ��
� ��
��
� ��
��
��
� ��
��
��
��
��
��
� ��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
�
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
�
��
��
�
��
��
�
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
278
��
Bulletin of Monetary, Economics and Banking, January 2014
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Lanjutan
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
�
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
�
�
�
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
IV. RESULT AND ANALYSIS
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
TFP growth, then its decomposition.
��first part of this chapter analyzes the
�
�� The
�� aggregate
��
��
�
��
In��
addition
to analyzing the entire period ��
(2000-2009),
we analyze tow period, 2000-2004
�
��
�
��
and 2005-2009. Analysis on aggregate industry will proceed to a more detail across subsector
�
��
��
��
of��
the ��
industry. The analysis on technical change
and efficiency change in sub sector level
is
��
��
��
carried
out by plotting them into four quadrants.
��
��
��
��
��
�� The
�� Industry
��
4.1.
Agregate TFP of Manufacture
��
��
��
��
Averagely the TFP grew by 7.44% per annum during 2000-2009. The major sources of
TFP growth was dominated by the growth of technical change, and then followed by economic
scale change, and finally the efficiency change. This result strongly indicate that during the
period of 2000-2009, the companies relied more on the use of new technologies and moved
toward optimal scale.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
The TFP growth slowed down in period 2005-2009 compared to the 2000-2004 period.
The source of TFP growth in 2000-2004 was efficiency change, while in 2005-2009 the main
source of TFP growth was technical change.
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
279
The strong efficiency change in 2000-2004 was associated with ongoing consolidation
after the financial crisis 1998. This included the improvement of investment climate to raise
the investor’s confidence. As the domestic demand and investment activity was weak, the
companies improved their productivity by increasing the production efficiency. The company
improved the use of intermediate input, improving production layout to shorten the switching
between work stations, aligning the workflow across workplaces (the concept of pull systems)
to reduce the accumulation of half finished product between work stations, and the application
of Lean Manufacturing conceptsto reduce the idle time between work stations. During 20002004, the slowed down technical change means the decline of production frontier, due to the
declining production capability of the machines. One possible reason is disturbance on machine
replacement as indicated by the low growth of investment (GFCF/PMTB) and the low realization
of investment (both FDI and domestic).
In contrast, during 2005-2009 period, technical change play greater role on TFP growth.
The average growth rate of investment and the realization of domestic and foreign investment
were higher compared to the previous period (Figure 18 and 19). The increase of aggregate
investment and the realization of domestic and foreign investment generally bring new
technologies.
During this period, though the technical change increased, the efficiency change or
catching up effects decreased. Similar studies on productivity in other countries explain the
reason for the decline of catching up effect when technical change increaseis the lack of human
resource capacity on adapting the new technologies.
The inability to catch-up indicates a low labor competency, either because of the low
education and low skill. This weakness may affect the ability of manufacturing sector to operate
optimally. In the long run, this may cause the foreign investment in Indonesia is limited to the
low technology one.
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
�
�
�
�
�
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
Figure 18.
The Average of PMTB Growth
Figure 19. The Average of foreign (PMA)
and Domestic (PMDN) Investment Growth
Bulletin of Monetary, Economics and Banking, January 2014
280
Unlike Indonesia, the increase of technical change in Malaysia, was not followed by a
decrease in efficiency change (Figure 20). The rank of Global Competitiveness Index during
2012-2013 for Indonesia was apparently far behind Malaysia, especially for the fourth pillar
(basic health and education) and fifth pillar (secondary education and training). This may explain
the difference of the TFP dynamics over the two economies (Table 5).
��
��
��
��
��
��
��
�
��
��
��
��
�
�
��
�
�
��
��
��
��
��
��
�
��
��
��
��
�
��
��
��
��
�
��
��
��
��
��
��
��
Figure 20. The Comparison of Technical Change and Efficiency
Change between Indonesia and Malaysia
��
��
�
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
�
�
��
��
��
��
�
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
�
��
��
�
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
��
��
��
��
��
��
��
��
��
��
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
281
4.2. TFP and Its Component Across Manufacturing Subsector
The model calculation show the total factor productivity of all subsectors grew during
2000-2009, except for Lamp industry (Table 6). Technical change became the major source
for the growth of TFP in most subsectors (about 75 percent of all 49 subsectors we observed).
On the other hand, efficiency change play greater role on the TFP growth in the following
subsectors: Spinning; Other foods, Leather goods, Footwear; Glass; Clay product; Building
and Construction; Equipment and Components of four or more wheel vehicles; and Other
Manufacturing.
Five subsectors with the highest TFP average growth during 2000-2009 are generally
classified in high-tech industry. They are Medical Equipment; Other Electrical Equipment; Electric
Motors and Engines; Heavy equipment;and Electricity and Telephone Cables. The source of
productivity growth for these subsectors is technical change.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
Bulletin of Monetary, Economics and Banking, January 2014
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
282
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
4.3.
Subsector Quadrant of Industry and Its Characteristics
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
We map subsector of industry into four quadrants based on their level of technical
��
��
��
��
efficiency change (positive
or negative).
��
�� Quadrant
�� I, includes
��
subsectors
with high technical change and positive efficiency
change.
This quadrant
be
��
��
��
�� should
��
��
��
��
for
high productivity sub sectors, and are considered as supreme
sub��
sector. ��
��
change
(high and low4), and their
��
��
��
��
��
��
Conversely, quadrant IV cover subsectors with low technical
change
growth
��
��
�� and negative
��
��
��
��
��
in��
efficiency change. This quadrant includes those subsectors��
with low productivity
with stagnant
��Quadrant II is for high technical change sector but
��with negative
��
��
��
progress.
efficiency
change.
��
��
��
��
��
This quadrant includes subsector industry with a low ability to catch up. Increasing number of
��
��
��
��
��
subsectors
in quadrant II will indicates the inability for the��
company��
to operate��
efficiently.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
4��
The cut point for technical change category is its median.
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
283
Our calculation shows that for the period of 2005-2009 (Table 8), the number of subsectors in first quadrant decline relative to previous period, 2000-2004 (Table 7). In contrast,
the number of sub-sectors in quadrant II increased across the two corresponding periods.
The decreasing number of industry sub-sectors in quadrant I and the growing number of
subsectors in quadrant II indicates a lack of development and innovation on managerial (working
procedures) along the production process. This affects the firm in two aspects, first, the ability
of manufacturing sector to operate at its potential level; and second, the ability of the labor to
adapt increasing technology.
The above condition is unfortunate considering the high technical change potentially
offer higher productivity. One way to overcome this problem is by developing the skill of the
workers to cope with higher technology.
284
Bulletin of Monetary, Economics and Banking, January 2014
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
285
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
286
Bulletin of Monetary, Economics and Banking, January 2014
The mapping of subsector onto the four quadrants may depend on the characteristic of
the firm within the subsector industry. Some variables gathered from the Survey of Medium
and Large Industry (SIBS) include: the intensity of research and development activities (R&D);
innovation; sales orientation; location of the company; the use of foreign investment facility;
type of ownership; and years of schooling.These variables are useful to explain our mapping;
though it requires formal modeling and statistical testing, which is beyond the scope of this
paper.
Table 9 below summarizes the characteristic of the firm within quadrant. We recall that
in terms of technical change and efficiency change, quadrant I is better than quadrant II and III,
and quadrant II or III is better than quadrant IV. However, apple to apple comparison between
quadrant II and III is not valid.
We see some interesting evident from the survey; first, the use of foreign investment
facility is associated with higher technical change and also higher efficiency change, which lead
them to quadrant I. Second, similar pattern apply when the company is partially or fully owned
by foreign. Third, if the company is located in industrial area, the likelihood to have better
infrastructure support increase. On the other hand the motivation for companies to learn from
each other is greater. These will help them increase their productivity; hence put them more
likely in quadrant I, II, then quadrant III.
��
��
��
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
�
�
�
�
�
�
�
�
�
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
��
The Dinamics of Total Factor Productivity Ofmedium and Large Manufacturing In Indonesia
287
V. CONCLUSION
This paper provides several interesting result. First the TFP of medium and large scale
companies in Indonesia grew 7.44% on average during the year of 2000-2009. The main
source of the TFP growth is technical change, followed by economic scale change, and finally
efficiency change.
During the period of 2000-2004, the source of TFP growth was efficiency change, while
for the period of 2005-2009, the source of TFP growth shifted into technical change, along
with the increasing investment activity.This is the second conclusion of this paper.
Third, even though the technical change increased, the catching-up capability (efficiency)
decreased in 2005-2009, showing the inability of the company to adapt the more advance
technology.
288
Bulletin of Monetary, Economics and Banking, January 2014
REFERENCES
Boediono, DR., “Seri Sinopsis Pengantar Ilmu Ekonomi No. 1 Ekonomi Mikro”, BPFE, edisi 2,
1999.
Coelli, Timothy J., et al. (2005), “An Introduction To Efficiency And Productivity Analysis”,
Springer Science & Business Media, Inc, edisi kedua.
Coelli T.J. (1996), “A Guide to DEAP Version 2.1: A Data Envelopment Analysis (Computer)
Program”, Centre for Efficiency and Productivity Analysis (CEPA), Department of Econometrics
University of New England, Australia.
Cooper, William W.; Seiford, Lawrence M. and Tone, Kaoru, (2007). “Data Envelopment
Analysis”, 2nd Edition. Springer.
Direktorat Evaluasi Kinerja Pembangunan Sektoral, (2010), “Perubahan Produktivitas Industri
Manufaktur Indonesia dan Faktor-faktor yang Mempengaruhinya: Analisis Panel Data 20002007”, Kementerian PPN/Bappenas.
Halim, Rizal Edy, (2010), “Marketing productivity and profitability of Indonesian public listed
manufacturing firms: An application of data envelopment analysis (DEA)”, Benchmarking:
An International Journal, Vol. 17 Iss: 6 pp. 842-857.
Kumbhakar, Subal C. and Lovell, C.A.Knox, (2004). “Stochastic Frontier Analysis”. Cambridge
University Press.
Ikhsan, Mohamad (2007), “Total Factor Productivity in Indonesian Manufacturing: A Stochastic
Frontier Approach”, Global Economic Review Vol. 36, No. 4, pp. 321-342.
Mahadevan, Renuka (2002), “A DEA Approach to Understanding the Productivity Growth of
Malaysia’s Manufacturing Industries”, Asia Pacific Journal of Management, 19, 587-600.
Pindyck, Robert S., dan Daniel L. Rubinfeld, “Microeconomics”, edisi ke-4, Prentice-Hall,
1998.
Prabowo, Handono E.T., dan Cabanda, Emilyn (2011), “Stochastic Frontier Analysis of Indonesian
Firm Efficiency: A Note”, International Journal of Banking and Finance, Vol. 8; Iss. 2, Article
5.
Saputra, Putu Mahardika Adi, (2011), “Analysis of Technical Efficiency of Indonesian
Manufacturing Industries: An Application of DEA”, International Research Journal of Finance
and Economics, Issue 66.
289
WRITING GUIDANCE
1. The paper should be original and should not violate any copyrights. The submitted paper
should have never been published or not being submitted to other publisher. The copyright
of the published paper is retained to the author.
2. The Bulletin of Monetary Economics and Banking provide a financial incentive between IDR
5,000,000.
3. The paper should be submitted in two formats (i) Microsoft Word (*.doc) and (ii) portable
digital file (*.PDF). These files should be sent to the following mail address:
[email protected] and Cc to: [email protected]
You may save your files in CD and send it to the following Editorial Address:
Bulletin of Monetary Economics and Banking
Department of Economic Research and Monetary Policy-Bank Indonesia
Building Syafruddin Prawiranegara, 20th Floor
JI. M. H. Thamrin No. 2 Central Jakarta, Indonesia
Ph. +62-21-2310108 / 2310408 ext. 4119, Fax: 62-21-3800394
4. To avoid missing fonts or other compatibility issues, any special characters or mathematical
expression (equations, symbols, matrix, etc.) must be written using Microsoft Equation.
5. The submitted paper should contain (i) an abstract of maximum one page A4, (ii) keywords
and (iii) JEL classification code. See the JEL code at http://www.aeaweb.org/journal/jel_class_
system.html.
6. The paper must contain the followings:
The background, the aim of the paper and its distinction to previous study
Theory and review of literatures
Methodology (quantitative methodology is preferred)
Result and analysis
Policy and further study implication
7. The citation should be in footnote and not in endnote.
8. The references must obey the following rule:
290
Bulletin of Monetary Economics and Banking, October 2010
a. Book:
Hanke John E. and Arthur G. Reitsch, (1940), Business Forecasting, Prentice-Hall, New
Jersey.
b. Article in journal:
Rangazas, Peter. (2000) “Schooling and Economic Growth: A King-Rebelo Experiment
with Human Capital”, Journal of Monetary Economics, October, 46(2), page. 397-416.
c. Article in book edited by other people:
Frankel, Jeffrey A. and Andrew K., Rose. (1995) “Empirical Research on Nominal
Exchange Rates”, in Gene Grossman and Kenneth Rogoff, eds.,”Handbook of International
Economics. Amsterdam: North-Holland, page. 397-416.
d. Working papers:
Kremer, Michael and Daniel, Chen. (2000) “Income Distribution Dynamics with
Endogenous Fertility”. National Bureau of Economic Research (Cambridge, MA) Working
Paper No.7530.
e. Mimeo or unpublished work:
Knowles, John. “Can Parental Decision Explain U.S. Income Inequality?”, Mimeo,
University of Pennsylvania, 1999.
f. Article from web or other electronic form:
Summers, Robert and Alan W., Heston. (1997) “Penn World Table, Version 5.6”
http://pwt.econ.unpenn.edu/
g. Article in newspaper, magazine or equal periodicals:
Begley, Sharon. (1993) “Killed by Kindness”, Newsweek, April 12, page. 50-56.
9. The paper should be submitted along with curriculum vitae complete with mail address and
phone number.