Demand for International Reserves in Turkey

Abstract

Having an important place in the international monetary system, international reserves held by central bank usually reflect country's economic strength in terms of international finance and trade. There are many reasons for holding international reserves by central banks such as financing the deficit in the balance of payment, managing the monetary and exchange rate policies, minimizing the negative effects of external shocks and reducing the cost of borrowing. Continuously changing and diversifying characteristics of these reasons affect the demand for reserves depending to the economic conditions of the country. Over the last ten years, there has been a tremendous increase in international reserves held by Turkish Central Bank. From 2002 to 2012, the reserves of the bank have risen from 20 billion dollars to 96 billion dollars, showing an increase more than four times. This sudden and huge increase in the foreign reserves drove us to determine and investigate the factors which induce the Turkish Central Bank to hold high level of reserves. Thus, the purpose of this study is to estimate and analyze the demand for international reserves held by central banks using the buffer stock model in the case of Turkey. The data used in the study is monthly and cover the period of 1990:03-2012:10. The buffer stock model was econometrically estimated by using the OLS method for three different models. Our findings indicate that the opportunity cost affected reserve demand much stronger than the reserve volatility in Turkish case.

Full text

The Romanian Economic Journal
63
Having an important place in the international monetary system, international
reserves held by central bank usually reflect country’s economic strength in terms of
international finance and trade. There are many reasons for holding international
reserves by central banks such as financing the deficit in the balance of payment,
managing the monetary and exchange rate policies, minimizing the negative effects
of external shocks and reducing the cost of borrowing. Continuously changing and
diversifying characteristics of these reasons affect the demand for reserves depending
to the economic conditions of the country. Over the last ten years, there has been a
tremendous increase in international reserves held by Turkish Central Bank. From
2002 to 2012, the reserves of the bank have risen from 20 billion dollars to 96
billion dollars, showing an increase more than four times. This sudden and huge
increase in the foreign reserves drove us to determine and investigate the factors
which induce the Turkish Central Bank to hold high level of reserves. Thus, the
purpose of this study is to estimate and analyze the demand for international
reserves held by central banks using the buffer stock model in the case of Turkey.
The data used in the study is monthly and cover the period of 1990:03-2012:10.
The buffer stock model was econometrically estimated by using the OLS method for
three different models. Our findings indicate that the opportunity cost affected
reserve demand much stronger than the reserve volatility in Turkish case.
1Yimaz Sinem , Karadeniz Technical University, e-mail: sinemyilmaz17@hotmail.com
Demand for International
Reserves in Turkey
Eyüboğlu, Sinem1
Yamak Nebiye2

The Romanian Economic Journal
64
Keywords: Central bank, Buffer stock model, Demand for international reserves,
International reserves.
JEL Classifications: E58 , F37
1. INTRODUCTION
The theory of demand for international reserves had been established
through important contribution of Heller (1966). In a few years the
theory has been extended by some studies (for instance Hamada and
Ueda (1977), Frenkel and Jovanovic (1981)). Since then, the demand
for international reserves has become one of the most remarkable
interests in policy and academic circle. This recent interest causes to
rapid increase in international reserves held by developing countries
such as Turkey. Over the past few decades, there has been a
tremendous increase in international reserves held by Turkish Central
Bank. For example, starting from a low level of U.S. $4.7 billion at the
end of March 1990, the stock of international reserves have increased
continuously to U.S. $18.8 billion by the end of December 2001 and
have reached U.S. $99.2 billion in end October 2012 (see Figure 1). In
other words, in twenty two years the growth rate of reserves has been
increased nearly 2010%.
Globally international reserves have grown very quickly especially in
the last decade. These growths of international reserves have been
even faster for developing economies than developed ones. There are
many reasons to hold international reserves by central banks. Those
reasons might be ordered as fixing the value of the national currency
change, supporting monetary and exchange rate policies, meeting
foreign debt, providing emergency protection and financing imports
respectively.
Despite being much benefits of holding reserves, if reserves hold
greater than a certain value, the country have more harm than good.
The optimal level of reserve which is determined by countries is
important to decide of the demanded amount of reserve. So, research

The Romanian Economic Journal
65
on optimal international reserves is important for sustainable
development.
Until present, there have been a great number of studies on the
demand for international reserves for developed and developing
countries. Generally Turkey was excluded for in panel data analysis
those developing countries dealing with the demand for international
reserves since there are still lack of studies on Turkey. In this paper we
intend to fill up this gap in the literature.
In fact, we analyze the demand for international reserves in the Turkey
context through using the buffer stock model of reserves. In their
recent cross-sectional time series analysis, Flood and Marion (2002)
asserted that the buffer stock model is so successful to define
international reserve holding for developing economies. The model is
based on the balance of two costs which are; the opportunity cost and
the adjustment cost. The opportunity cost of holding reserves is the
difference between return on alternative investment and on reserves.
The adjustment cost is generally related to policy action lead to welfare
lost.
The rest of the paper is organized as the review of empirical literature
about the particular issue in section 2, trends in international reserves
in Turkey in section 3, the model of the demand for reserves in
section 4, the results of the empirical research in section 5, and
conclusion remarks of the study in the last section.
2. REVIEW OF THE EMPIRICAL LITERATURE
This section has reviewed some empirical studies on the demand for
international reserves from the period 1981 to 2009. Empirical studies
can be separated broadly into two groups. The first group has
determined the demand for international reserve for cross countries
(see Table 1). Most of cross country studies group countries into
developed and developing countries. The second group has
investigated the demand for international reserve under the time series
data for an individual country (see Table 2).

The Romanian Economic Journal
66
Result of Frankel’s (1981) study is consistent with the predictions of
the buffer stock model. Moreover Flood and Marion (2002) have
indicated that the buffer stock model worked well in the era of high
capital mobility and country characteristics could affect the cost of
adjustment. On the other hand, Aizenman and Marion (2002) have
found that sovereign risk and costly tax collection led to hold high
level of reserves to cover fiscal liabilities for developing countries.
Cheung and Ito (2009) have concluded that developing economies
tend to hold higher level of reserves than developed ones.
Table1
Cross-Country Studies on the Demand for International Reserve
Author
Country
Period
and data
frequency
Dependent
variable
Explanatory
variable
Methodology
Frankel
(1981)
For 22
developing
countries
1971-1975
annual
data
R
( Level of
reserves)
Opportunity
cost, adjustment
cost, import
OLS
Aizenman
and
Marion
(2002)
For 122
developing
countries
1980-2002
annual
data
R/P
(deflated by
U.S GDP
deflator)
GDP,
Population,
Volatility of real
export receipts,
volatility of
nominal
effective
exchange rate,
share of imports
of goods and
services in GDP
Panel
Regression
Flood and
Marion
(2002)
For 35
countries
1988-1999
annual
data
R, R/M,
R/P,
R/GNP or
R/M2
Volatility of
Reserves,
opportunity
cost, Nominal
effect exchange
rate volatility,
GMM

The Romanian Economic Journal
67
Elbadawi (1988) has concluded that the variability measure was
positive and significant in Sudan to accumulate reserves. Ford and
Huang (1994) have indicated that domestic monetary disequilibrium
significantly affected reserve demand in China. Ramachandran (2004)
has found that the opportunity cost played a greater role than reserve
volatility to determine the level of reserves in India. Parallel with
Ramachandran’s (2004) framework for Pakistan, Jalil and Bokhari
(2008) have found similar results through using the same method.
However Prabheesh, Malathy and Madhumathi’s findings (2007) have
showed that capital account vulnerability is more sensitive than its
opportunity cost for reserve accumulation in India. Similarly Ra (2007)
has showed that after crises the Korean reserve demands have
becomed more sensitive to the adjustment cost and the openness than
opportunity cost. Sehgal and Sharma (2008) have found that risky
capital flows and exchange rate volatility have positive impact on
international reserve demand. And export growth is found to be
significant. Nor, Azali and Law (2008) have concluded that the current
account balance and short term external debt have significant impacts
on the demand for international reserves in Malaysia.
ratio of export
plus import to
GDP, ratio of
gross capital
flows to GDP
Cheung
and Ito
(2009)
For 119
countries
1975-2004
annual
data
R/GDP
Traditional
macroeconomic
variables,
financial
variables,
institutional
variables,
dummy
variables
Panel
Regression

The Romanian Economic Journal
68
Table 2
Individual Country Studies on the Demand for International
Reserve
Author
Countr
y
Period
and data
frequenc
y
Dependen
t variable
Explanatory
variable
Methodology
Elbadawi
(1988)
Sudan
1971-1982
quarterly
data
First
difference of
reserves
import to
GDP ratio,
variability of
reserves,
uncertainty of
reserves, scale
variables,
remittances
ECM
Ford and
Huang (1994)
China
1956-1989
annual data
First
difference of
reserves
Sum of the
industrial and
agricaltural
output,
uncertainty of
reserves,
import to
GDP ratio
ECM
Ramachandra
n (2004)
India
1999-2003
weekly data
Level of
reserves
Adjustment
cost,
opportunity
cost
ARCH/GARC
H
Prabheesh,
Malathy and
Madhumathi
(2007)
India
1983-2005
monthly
data
Level of
reserves
Economic
size, current
account
vulnerability,
capital
account
vulnerability,
exchange rate
flexibility,
opportunity
cost
VECM

The Romanian Economic Journal
69
Ra (2007)
Korea
1973-2005
1990-1997
1998-2005
monthly
data
Level of
reserves
Adjustment
cost,
opportunity
cost,
openness,
scale variables
Johansen and
Juselius (1992)
cointegration,
ECM
Sehgal and
Sharma (2008)
India
1990-2006
monthly
data
Level of
reserves
GDP, import
to GDP ratio,
short run
external debt
to GDP ratio,
portfolio
investment to
GDP ratio,
opportunity
cost, export
growths
VECM
Jalil and
Bokhari (2008)
Pakistan
1995-2005
monthly
data
Level of
reserves
Adjustment
cost,
opportunity
cost, dummy
for regime
change
ARCH/GARC
H
Nor, Azali and
Law (2008)
Malaysia
1970-2004
annual data
Level of
reserves
Scale
variables,
variability
measurements
, propensity to
import,
current
account
balance, real
short term
external debt
ARDL
3. TRENDS IN INTERNATIONAL RESERVES IN TURKEY
The level of Turkey’s international reserves has showed massive
growth in the period from 1990 to 2010. Figure 1 exhibits significant
increasing trend in international reserves. Especially, over the last ten
years, there has been a tremendous increase in international reserves

The Romanian Economic Journal
70
held by Turkish Central Bank. From 2002 to 2012, the reserves of the
bank have risen from 20 billion dollars to 96 billion dollars, showing
an increase more than four times. The highest growth rate of reserves
in Turkey has become in January 1995 (35%).
Also Turkish Central Bank reserves have been continuously increasing
since 1987s. In October 1990 the reserves increased from U.S. $6
billion to U.S. $26 billion in November 2002. In October 2008 the
amount climbed to U.S. $72 billion and finally they reached U.S. $99
billion in October 2012.
Figure 1
Turkey’s International Reserve (Million US Dollar)
Source: Central Bank of the Republic of Turkey (CBRT).

The Romanian Economic Journal
71
4. THE MODEL OF THE DEMAND FOR
INTERNATIONAL RESERVES
In this section we used the buffer stock model to explain optimal
international reserve movements. Heller (1966) determined the
optimal stock of reserves in terms of a rational optimising decision
that including equating the marginal cost and benefit of holding
reserves. To determine the optimal stock for reserves, Frenkel and
Jovanovic (1981) expanded Heller’s framework and developed a
theoretical model which is based on the principles of inventory
management.
The buffer stock of Frenkel and Jovanovic (1981) characterized
reserve movements in continuous time period as a Wiener process2, by
the following stochastic equation:
dR (t) = -µdt + ódW (t) (1)
where R (t) is reserve held in time t and W (t) is the standard Wiener
process with mean zero and variance t. In variably the distribution of
reserve holdings R (t) can be defined as:
R (t) = R* - µt + óW (t) (2)
where R* is the optimal stock of reserves, µ is deterministic part of
sudden change in reserves and ó is the standard deviation of the
Wiener increment in reserves. The estimating reserve demand
equation can be written as:
Log Rt = b0+b1lnót+b2lnrt+ut (3)
where b0,ó and r are fixed, adjustment and opportunity cost of holding
reserves respectively and ut is white noise error. The theoretical
prediction of the model for the parameters are that b1= 0,5 and b2= -
2 The Wiener Process is a continuous time of a simple random walk with independent
increments (Jalil and Bokhari (2008), 40).

The Romanian Economic Journal
72
0,25. Substantially equation (3) identifies two macroeconomic costs for
determining the optimal stock of reserves. The first is the adjustment
cost (ót), which is based on the frequency of adjustment. The second
is the opportunity cost (rt), which is incurred the forgone earnings on
reserve accumulation.
5. EMPRICAL ANALYSIS
We estimate equation (3) using monthly data for the period from
1990:03 to 2012:10. The data on reserves and one month forward
interest rate are obtained from the Central Bank of the Republic of
Turkey. To be consistent with the literature on reserves for measuring
the adjustment cost we is defined rolling standard deviation of change
in reserves and opportunity cost which is defined by using one month
forward interest rate.
Firstly we estimate equation (3) for three different adjustment costs
which are defined as 4 month, 6 month and 8 month rolling standard
deviation of change in reserves. The results are shown in Table 3
respectively.
Table 3
Results of Buffer Stock Models for Three Different Adjustment
Cost
Indicators
Model 1
Model 2
Model 3
bo
9.7575***
(0.5979)
8.4879***
(0.6533)
7.7133***
(0.7765)
b1
0.5204***
(0.0673)
0.6609***
(0.0760)
0.7461***
(0.0885)
b2
-
0.8950***
(0.0806)
-
0.8314***
(0.0759)
-
0.7916***
(0.0769)

The Romanian Economic Journal
73
R2
0.7392
0.7700
0.7906
F- statistic
375.7310
440.415
429.9950
D-W
0.3218
0.2613
0.2412
Parenthesis indicates Newey-West standard
errors. ***, indicates that coefficients are
significant at 1% level.
As seen in Table 3 the buffer stock model explains averagely 76,66%
of the reserve demand. All the variables signs in all regression are
consistent with the theoretical predictions. Furthermore, for three
models the volatility coefficients are found to be higher than the
theoretical prediction. All types coefficient of volatility are positive and
significant at 1% level. This means that volatility increases by 1% the
reserves increases by 0.5204%, 0.6609% and 0.7461% for three
models respectively. However, model 1 is the closest model to the
theory among all three models 3.
Further Figure 2 shows that the behavior of observed and estimated
values of reserves gained from equation (3). During March 1990 to
March 1996, the CBRT hold reserves less than optimal level averagely
3 Additionally we measure the volatility by modelling the variance of reserve changes
using ARCH specification. However the results of ARCH model worse than all three
models because of the coefficients similarity with theoretical prediction.
Please do not put here your name or authors names.
You can insert your equations
with Latex formulas surrounded by \( and \) like this
∑ n i=1 x 2 i
or
surrounded by two dollar signs like this
∫ π 0 sinxdx=2
or
you can save your equations on your desktop as simple images, attach them in the
Figures field and insert them in the Full manuscript text.

The Romanian Economic Journal
74
U.S. $7 billion. When we compare to results of 2012 year, CBRT hold
averagely U.S. $93 billion of optimal reserves. But last two months
estimation results show that the level of optimal reserves has reduced
U.S. $77 and U.S. $64 respectively. Looking at these conclusions, we
can say that recent two months reserve movements express the
optimal behavior more than the buffer stock model.
Figure 2
Observed and Estimated Values of International Reserve
Demand Equation
6. CONCLUSION
The current paper has estimated the buffer stock model using an
analysis of time series for Turkey from 1990 to 2012 using monthly
data. All of the estimates of buffer stock models indicate that the
opportunity cost affected determining the level of reserves much
stronger than the reserve volatility but not to the significant effect for

The Romanian Economic Journal
75
model 3. This result is consistent with Ramachandran (2004) and Jalil
and Bokhari (2008). As said by Ramachandran (2004) this can perhaps
be attributed to the fact that capital outflow in Turkey as in India and
Pakistan is less free than capital inflow. Furthermore, in the second
quarter of 2012 Turkey’s total foreign debt stocks have declared U.S.
$323 million by CBRT. Maybe to say that a large part of reserve
accumulation is derived from the level of foreign debt would not be
wrong.
REFERENCES
Aizenman, J. and Marion, N. (2002) ‘The High Demand for
Internatıonal Reserves in the Far East: What’s Going On?’. Working
Paper number 9266.
Cheugn, Y. W. and Ito, H. (2009) ‘A Cross-Country Analysis of
International Reserves’, International Economic Journal, vol. 23, no. 4,
pp. 447-481.
Elbadawi, I. A. (1988) ‘The Sudan Demand for International Reserve:
A Case of a Labor Exporting Country’, Economica, vol. 57, pp. 73-89.
Flood, R. and Marion, N. (2002) ‘Holding International Reserves in an
Era of High Capital Mobility’, International Monetary Fund Working
Paper number 02/62.
Ford, J. L. and Huang, G. (1994) ‘The Demand for International
Reserves in China: An ECM Model with Domestic Monetary
Disequilibrium’, Economica, vol 67, pp. 379-397.
Frenkel, J.A. (1981) ‘Optimal International Reserves: A Stochastic
Framework’, The Economic Journal, vol. 91, no. 362, pp. 507-514.
Gray, S. (2011) ‘Central Bank Balances and Reserve Requirements’,
IMF Working Paper, 1-55.
Hamada, K. and Ueda, K. (1977) ‘Random Walks and The Theory of
Optimal International Reserves’, The Economic Journal, vol. 87, pp.
722-742.

The Romanian Economic Journal
76
Jalil, A. and Bokhari S. (2008) ‘The Optimal Demand for Foreign
Exchange Reserves in Pakistan’, International Journal of Applied
Econometrics and Quantitative Studies, vol. 5, no. 1, pp. 33-48.
Nor, E., Azali, M. and Law, S. H. (2008) ‘Internatıonal Reserves,
Current Account Imbalance and External Debt: Evıdence from
Malaysıa’, Journal of Economics and Management, vol.16 no. 1, pp.
47-76.
Prabheesh, K. P., Malathy, D. and Madhumathi, R. (2007) ‘Demand
for Foreign Exchange Reserves in India: A Co-integration Approach’,
South Asian Journal of Management, vol. 14 no.2, pp. 36-46.
Ra, Hee-Ryang (2007) ‘Demand for International Reserves: A Case
Study for Korea’, The Journal of the Korean Economy, vol. 8 no. 1,
pp. 147-175.
Ramachandran, M. (2004) ‘The Optimal Level of International
Reserves: Evidence for India’. Economics Letters, vol. 83, pp. 365-
370.
Sehgal, S., Sharma C. (2008) ‘A Study of Adequacy, Cost and
Determinants of International
Reserves in India’, International Research Journal of Finance and
Economics, vol. 20, pp.75-90.