Volatility in the Call Money Rate and Efficacy of Monetary Policy Operations in Bangladesh

Policy Analysis Unit (PAU)





Working Paper Series No: WP 0802






Volatility in the Call Money Rate and Efficacy of

Monetary Policy Operations in Bangladesh




Md.
Shahiduzzaman



Mahmud Salahuddin Naser




November 2007


Policy Analysis Unit (PAU)
Bangladesh Bank
Head Office, Dhaka, Bangladesh
(www.bangladeshbank.org.bd)
(www.bangladesh-bank.org)

---

Policy Analysis Unit* (PAU)



Working Paper Series No: WP 0802





Volatility in the Call Money Rate and Efficacy of

Monetary Policy Operations in Bangladesh




Md. Shahiduzzaman
Research Economist, Policy Analysis Unit
Bangladesh Bank

Mahmud Salahuddin Naser
Joint Director, Monetary Policy Department
Bangladesh Bank


November 2007

Copyright © 2007 by Bangladesh Bank

* In an attempt to upgrade the capacity for research and policy analysis at Bangladesh Bank
(BB), PAU prepares and publishes several Working Papers on macroeconomic research every
quarter. These papers reflect research in progress, and as such comments are most welcome. It is
anticipated that a majority of these papers will eventually be published in learned journals after
the due review process. Neither the Board of Directors nor the management of, or any agency of
the Government of Bangladesh, however, necessarily endorses any or all of the views expressed
in these Papers. The latter reflect views based on professional analysis carried out by the research
staff of Bangladesh Bank, and hence the usual caveat as to the veracity of research reports
applies.

1

---

Volatility in the Call Money Rate and Efficacy of Monetary
Policy Operations in Bangladesh





∗
Md. Shahiduzzaman
Mahmud Salahuddin Naser*



November 2007



Abstract
This paper examines the pattern of volatility in the call money rate and its association to monetary
policy operations of Bangladesh Bank (BB). In the process, it explores empirical issues like the co-
movement of money market rates and the presence of heteroscedasticity in the call money rate. The
first stage analysis supports the view that money market rates co-move and there is a co-integration
relationship between overnight money market rate and BB's policy interest rates. At the second stage,
long-run properties of the data were incorporated while trying to investigate the efficacy of
operational policy of BB in the indirect policy regime by separating the short-run and long-run
effects. As the variances of the error terms are not constant,, OLS estimation provides a false sense of
precision. A GARCH model is therefore estimated which suggests that the operational policy of BB is
effective in influencing the rate volatility in Bangladesh. The empirical evidence of this study will
assist in better understanding the interest rate channel of monetary transmission, overnight money
market behavior, and operational policy of BB.


Key words: Call money rate, volatility, policy rate, co-movement

JEL classification: G21, E52, E58




∗ The views expressed in this paper are those of the authors and in no way reflect the views or opinion of the
Bangladesh Bank. The authors are grateful to Professor Syed M. Ahsan, Former WBI Resident Economic
Adviser, Bangladesh Bank for giving valuable comments on a earlier version of the paper and Dr. Mustafa K.
Mujeri, Chief Economist for comments on the final draft. Comments and suggestions are welcome by the
authors and may be forwarded to mszaman_bb@yahoo.com.

---

1. Introduction
With economic reforms and financial liberalization, Bangladesh started to put greater
reliance on a market based financial system since the early 1990s. Overtime, among others,
administered interest rate structures have been abolished and deregulation in the banking
sector has taken place. Accordingly, Bangladesh Bank (BB), the central bank of the country,
started reinforcing indirect monetary policy operations. The transition to this framework has
been accompanied by some indirect monetary policy arrangements, namely repo (repurchase
agreement), reverse repo and interbank repo from July 2002, April 2003, and July 2003
respectively. Bangladesh also entered into a floating exchange rate system in May 2003. All
these developments have generated a profound influence over the functioning and
development of the country’s money market. Through time, an increasing focus has also
been given to monitor the day-to-day activities in the overnight money market (call money
market) in order to pursue a prudent monetary policy by the central bank.2
There are various reasons why central banks around the world pay greater attention to the
stability of the call money market. It is generally viewed that the uncertainty caused by
volatility in the call money rate3 gives confusing signals to the market participants about the
stance of the monetary policy. In addition, as other interest rates in the money market come
with different lags, it is the call money rate, which the central bankers and market
participants can monitor on a daily basis. It is also believed that volatility in the overnight
money market can be radiated out to other money market rates as these rates reflect the
underlying changes in marginal cost of borrowing.
In a market where overnight rates fluctuate from a very high to a very low end, market
players, especially banks, are encouraged to keep higher excess reserve than under normal
situation to meet the uncertainties and can even be reluctant to purchase near cash assets like
bills and bonds from the primary market. The situation exacerbates if there is no active
secondary market for bills and bonds where banks can easily convert those to cash
reasonably quickly. In such a situation, open market operations (OMOs) conducted by the
central bank may not be such effective as to influence the excess liquidity in the market.
Volatility can also be transmitted to foreign exchange market; the difficulty specially arises
when both the call money and foreign exchange markets are in pressure. In case of shortage
of liquidity and/or higher demand for liquidity, central bank usually injects money to
minimize the deviation of call rate from its policy interest rates. In such a situation, the
injection of liquidity may in turn increase the demand side pressure in the foreign exchange
market, creating a possibility of the exchange rate to depreciate further. In Bangladesh, a
situation like this happened in the second half of FY06, when the weighted average daily call
money rate varied between 7.07 percent and 40.37 percent and the monthly average call rate
ranged between 10.84 percent and 21.54 percent. However, this was the period when foreign
exchange market was also under pressure; BB found difficulties in injecting fresh money for
the fear of deterioration in the domestic currency value.

2 Theoretically, call money market is the institutional arrangement where short-term funds flow from banks and
other financial institutions holding excess liquidity to banks and financial institutions in need of immediate
funds without any collateral security. Call money market is synonymously used as overnight money market in
different occasions.
3 Call money rate is the interest rate, which is determined in the call money market. Throughout the paper, call
money rate is synonymously used as call rate, overnight money market rate, and overnight rate.

3

---

Call money rate for overnight transactions reflects the demand for and supply of liquidity in
the market. Therefore, the use of monetary policy instruments to influence the liquidity
position in the market places a substantial impact on the rate. A common objective of the
central bank's monetary policy is thus to minimize the persistent deviations of the overnight
money market rates. It is therefore expected day-to-day implementation of monetary policy is
linked to the stability of the overnight money market rate. On the other side, if market
participants find that the central bank's operational measures can effectively control the
movement of market rates then they incorporate the belief into their demand. Obviously, the
success of this process depends on the effectiveness of the transmission channel of the
interest rates. However, if such success holds, credibility of the central bank would be
enhanced, which further helps policy measures to become more effective. It is also argued
that central banks can use open mouth operations (signaling) rather than any actual
interventions to keep market rates at a target level, if such credibility holds (Guthrie and
Wright 2000). Some studies also suggest that interest rates tend to move together, especially
in the long run, and there is a tendency that market-clearing rates converge towards the
central banks' rate for open market operations.
Given the above reasons, there is a common objective of a central bank's operational policy
to minimize the persistent deviations of the overnight money market rate(s) from its policy
rates. This study investigates whether the choice of operational framework of monetary
policy has been related to the pattern of call money rate in Bangladesh. In order to do so, we
first examine the long-run relationship between call money rate and 28-day Treasury bill (T-
bill) rate. We have chosen these two rates because of their importance and visibility in the
short run money market in the country. Call money rate is the market-clearing rate and an
information variable in the process of implementing monetary policy while 28-day T-bill rate
is the default risk free rate in the Bangladesh money market. Having investigated the long-
run relationship between these two important rates, we then build up an empirical model that
captures the feature of the data, in particular the time-varying nature of volatility, in order to
investigate the efficacy of various monetary policy arrangements to stabilize the volatility in
the call money rate. The empirical results presented in the paper have important implications
for monetary policy, particularly with respect to the interest rate channel of monetary
transmission.
The outline of the paper is as follows. Following the introduction in section 1, section 2
reviews some previous studies in this area in the context of Bangladesh and some other
countries. Section 3 discusses the practices of monetary policy operations and various factors
influencing volatility in the call money rate. Section 4 enumerates some stylized facts and
statistical characteristics of the data and section 5 incorporates the empirical analysis. Finally,
section 6 offers the conclusions and recommendations.
2. Review of Previous Studies
In recent decades, a significant body of empirical research has emerged on the behavior of
the call money market and its relationship with monetary policy in the context of developed
countries; the empirical research on the issue is however scarce in the context of developing
countries like Bangladesh. This is probably because of the underdevelopment of the money
market or non-following an interest rate targeting for monetary policy operations. In
addition, unlike monetary target, interest rate channel is not well understood. Nevertheless,

4

---

movements in the interest rates have a profound impact on businesses and economic
activities in any country, therefore studies in this regard can bring out important policy
implications.
In Bangladesh, few studies that so far have been done on the issue are descriptive in nature
and ignore various statistical properties of data. The studies, in general, attempted to study
the sources and impacts of fluctuations in the call money rate and implications of the trend of
the market behavior (see, for example, Sarker 1999; Akhtaruzzaman, Mahfuza and
Masuduzzman 2005). These studies, however, enumerate some important observations and
recommendations to facilitate further development of call money market and to limit
fluctuations. These include, enforcing indicative limits on individual banks to restrict their
exposure in inter-bank transactions, which draw funds disproportionately to their capital or
deposit base.
Sarno and Thornton (2002) investigate the dynamic relationship between two important
short-term interest rates in US financial markets-the federal funds rate and the 3-month T-bill
rate using a general non-linear asymmetric vector equilibrium correlation model (ECM).
Regardless of whether the argument that federal funds rate and T-bill rate move together
because they are linked as the expectations hypothesis (EH)4 holds, finding of their empirical
results provide strong evidence of a co-integrating relationship between the federal funds rate
and the T-bill rate.
Wetherilt (2002) examines whether the choice of policy instruments and other reforms to
Bank of England's money market operations affect the Bank's objective of minimizing
persistent deviations of the relevant money market rates(s) from its policy rate using daily
money market rates, ranging from the overnight to twelve-month maturity. The author
developed an empirical modeling framework, namely a single equation component GARCH
model, to analyze the dynamic behavior of overnight and other short-term rates and a vector
error correction model (VECM) to analyze the dynamic interactions between the entire
spectrum of short-term rates. This model captures the time varying nature of volatility of the
key money market rates i.e., the overnight and two-week rate. The study demonstrates that
higher spreads between the two-week market rate and the official repo rate result in lower
money market volatility and rate dynamics at the short end of the money market curve, and
hence, the effects at the longer end are much weaker.
Palombini (2003) found that the in the Italian interbank market, largest increase in volatility
and the most notable variations of its intraday pattern occur at the last working day of reserve
maintenance period and at the end of each quarter. Furthermore, he found that overnight
interest rate volatility is not influenced by trading volume. This finding indicates the
difference between a financial market, where interest rate level is determined by information
arrival and the market for overnight liquidity, where the volume of trading is more influenced
by institutional factors like the functioning of the payment system.

Joshi (2004) assesses the volatility pattern of the call money rate in India to estimate its
sensitivity vis-à-vis the Reserve Bank of India's liquidity adjustment facility (LAF) decisions.

4 According to the expectation hypothesis of the term structure of interest rate, it is assumed that the T-bill rate
is equal to the market's expectation for the federal funds rate over the term of the T-bill rate plus a risk
premium.

5

---

Moreover, he attempted to show how regulatory changes related to other instruments in the
money market might have affected the functioning of the inter-bank money market. He used
Nelson Beveridge (NB) time series decomposition and an ARCH-M [1,1] model. The NB
decomposition is used to differentiate between permanent and cyclical components in a time
series while ARCH imposes a systematic structure to the variance process making it
amenable to interpretation and use in forecasting. The evidence provided by the empirical
model brings out that while the call money rate is tracked reasonably accurately during
surplus liquidity conditions, the predictive power suffers a loss when liquidity shortage
suddenly emerges. In addition, it argues that introducing a wider range of eligible collateral
in the repo market could help in improving the efficiency of interest rate targeting.
The above review provides some empirical evidence toward understanding the behavior of
overnight money market rates, and implications of monetary policy measures in maintaining
the rate stability. Two important findings are noteworthy in view of the present study. One is
that money market rates are interlinked and there exit long-run relationships between money
market rates and policy rate of the central bank. The other is that recent arrangements of the
indirect monetary policy operations place significant impact of the movement of overnight
money market rates. In addition, identifying and capturing the time varying nature of
volatility in the empirical models is an important lesson for the methodology part of this
study.
3. Overnight Money Market and Monetary Policy Operations in Bangladesh
The large swings in the overnight call money rates are usually not isolated occurrences, but
are stemmed from overall economic activity and conduct of day-to-day monetary policy
operations by the central bank. In influencing interest rates, BB has a choice of instruments
grouped under three headings: Open Market Operations (OMOs), discount window, and
reserve requirements. The procedures that BB employs in using these instruments to
implement monetary policy affect the demand and supply conditions in the inter-bank market
for short-term funds, which forms part of money market named market for bank reserves,
where banks and other financial institutions trade in their reserve balances at BB to meet
intraday liquidity needs.5 The call money rate arises, in large part, as an interaction of the
demand for and supply of bank reserves jointly determined by the optimizing behavior of the
money market participants and of the BB and the efficient distribution of reserves throughout
the banking system.
Banks settle transactions among each other on the books of the central bank. The bank
reserves-banks balances at the central bank- are the ultimate means of settlement. Therefore,
banks have a demand for reserve balances that arise firstly from the demand for required
reserve in order to fulfill legal reserve requirement, the portion of reserve balance that BB
requires the banks to maintain.6 Secondly, demand for reserve balances arises from demand

5 In the rest of this article the term ‘banks’ will generally be used to cover both banks and other financial
institutions who take time deposits and are required to maintain a certain portion of their deposit liabilities as
statutory reserve.
6 The BB regulation requires that the banks and non-bank financial institutions licensed from the Bank be
required to maintain a certain percent of their deposit liabilities as statutory reserve. Presently BB requires the
scheduled banks to maintain 18 percent of their deposits as Statutory Liquidity Ratio (SLR), where Cash
Reserve Requirement (CRR) must be kept at the rate of 5 percent on daily average on two-week basis but no
less than 4 percent on any day. Non-bank financial institutions who take term deposits are required to maintain

6

---

for excess reserve. Excess reserve is the portion of reserve-balance that banks willingly keep
in order to settle transactions among themselves and between their customers, which takes
place through the transfers from the account of one bank at BB to that of another bank. The
demand for bank reserves affected by opportunity cost of holding reserves i.e. market interest
rate, expectations of future market interest rate, and the cost of avoiding default.
A required reserve is calculated as a proportion of average level of deposits held over a
month, called the reserve computation period. The calculated amount of reserve requirement
must be satisfied on average over a two-week period, called reserve maintenance period. BB,
like most other central banks, builds a lag between the computation period and the
maintenance period where the reserve maintenance period follows the reserve calculation
period. The lagged reserve accounting system helps the banks to estimate accurately their
reserve demand for a maintenance period.
From August 2004, BB allowed reserve-averaging system within a maintenance period with
a limitation that the daily reserve balance must not remain below a minimum requirement
level.7 The reserve averaging provision in bank reserves permits individual banks to enjoy
considerable flexibility to manage their daily accounts and to use part of their reserves to
offset short-term or seasonal fluctuations in liquidity. This flexibility in turn provides greater
interest rate sensitivity in the demand for bank reserves during the early and the middle part
of the maintenance period. Because required balance is met on a two-week daily average
basis, banks can substitute balance across days. For example, if a bank finds itself with
balance above the average requirement level on a particular day, it can offset the surplus
balance by holding lower balances (but not below the minimum levels) on subsequent days in
reserve maintenance period. Similarly, if a bank finds itself with balance below the average
requirement level on a particular day, it can make it up on subsequent days in the reserve
maintenance period. This ability to adjust reserve holdings across a maintenance period eases
the pressure on banks to borrow or lend on a single day, reducing the possibility that supply
and demand forces will cause call money rate to fluctuate abruptly.
Banks, however, would be indifferent about the amount of reserves they hold if they expect
no significant fluctuations in the inter bank rate. Moreover, they may be cautious to deviate
from the average level of reserve position they maintain for the fear of not being able to
adjust their position as the number of days remaining in the maintenance period reduces. As a
result, the demand for bank reserves tends to be interest inelastic especially at the end of the
reserve maintenance period. In contrast, if the banks are maintaining a higher balance
requirement, they would face less risk of an end-of day balance below required level and
have more flexibility to postpone purchases of funds within the maintenance period. If
enough banks are in low reserve positions or when payment flows in the banking system are
heavy because of seasonality, a distributional imbalance of funds among banks or an
aggregate surplus or shortage of funds is more likely to trigger volatility in the call money
rate (Bennell and Hilton 1997).

rate of 5 percent of total liabilities as SLR, of which 2.54 percent of term deposit must be kept as CRR on daily
average on bi-weekly basis but no less than 2 percent on a day.
7 According to reserve averaging provision, banks are required to keep with Bangladesh Bank cash reserve
requirement daily at the rate of 5 percent of their average deposit liabilities on two-week basis provided that the
amount of cash reserve would not be less than 4 percent in any day effective from 10 October 2005.

7

---

With lagged reserve accounting, total reserve in a given period-the bulk of which are
required reserve-are essentially predetermined by the level of deposits one month earlier. As
changes in demand for bank reserves arise mainly from changes in reserve requirement,
hence demand for bank reserves does not fluctuate considerably over a shorter period.
Therefore, it is the supply factors, which mainly affect the equilibrium interest rate in the
market over a short run. As the supply factors are largely influenced by various monetary
policy operations including OMOs, central banks do have enough control over the movement
of the equilibrium interest rates in the market. Therefore, the movements of the interest rates
are strongly linked with the daily conduct of monetary policy operations by central bank in
the short end.
The supply of bank reserves in a single day is determined by BB's holding of government
securities, borrowing by banks from BB, and autonomous supply of bank reserves, if no
OMOs are conducted. Some of the transactions causing liquidity flows are pre-known to BB;
these include, among others, holding of government. securities i.e. repo, reverse repo and T-
bills etc. However, transactions like changes in currency demand and transfer of public funds
to and from BB are referred to as autonomous supply, which is not pre-known by BB at the
time of conducting OMOs.
In addition to sudden autonomous changes in liquidity, some seasonal factors cause
significant day-to-day variations in the central bank's balance sheet, disrupt the central bank’s
management of liquidity, and lead to lower degree of precision of liquidity projections. As a
result, in anticipating the supply condition, projection of autonomous flow of liquidity
(autonomous supply of bank reserves) arising from various transaction of BB with the rest of
the economic system is very important in determining the market clearing rates.
In implementing daily OMOs, BB attempts to ensure that the supply of balances satisfies
demand at a rate consistent with its policy interest rates. BB makes estimates of the changes
in factors affecting its balance sheet and causing autonomous supply of bank reserves in
order to project the quantity of reserves that would be available before conducting any
OMOs. BB at the same time predicts demand for bank reserves to achieve its operational
objective. The excess reserves prior to BBs liquidity operations could then be estimated ex
ante by taking differences in demand and supply of bank reserves. BB considers these
differences in order to assess the required scale of its intervention in the money market
through daily OMOs.
If the demand for bank reserves exceeds the supply, BB judges that it needs to make more
reserves available and it will add reserves through open market purchase i.e. lending more
that the counterparties have to repay in OMOs maintaining on that day. If demand is less than
supply of bank reserves, BB drains reserves through an open market sale i.e. lending less
than the amount the counterparties are repaying. BB thus offsets the deviations by injecting
(withdrawing) liquidity into the banking system in an equivalent amount and thus bring the
amount of reserve money expansion in line with its target so that supply of broad money can
be kept under control. Therefore, while BB with its OMOs directly affects the amount of
liquidity in the market by shifting the supply curve, it can also affect the slope of the demand
curve by changing the reserve requirement. Through reserve requirements and conducting
appropriate OMOs, BB affects demand for or supply of bank reserves or both, thereby
limiting the fluctuations in the equilibrium rate in the market.

8

---

4. Data and Stylized Facts
Figure1 portrays the movement of interbank call money rate and the 28-day T-bill rate for the
last one decade. The figures reveal that the call rates have been highly volatile since 2002 as
compared with the historical average followed by an unprecedented stability in FY07. Table
1 shows that the average call rate and standard deviation were highest in 2006. Weighted
average call rate during the last ten years varied from 4.25 percent to 21.54 percent. On a
yearly basis, starting from 8.66 percent in 1997, the mean call rate dropped in the following
three consecutive years and reached at 7.31 percent in 2000, then rose again to 9.18 percent
in 2002 before touching the bottom at 5.91 percent in 2004-the lowest ever in the entire
period. The call rate again rose sharply to 11.10 percent in 2006 along with substantial
increase in volatility.
Figure 1: Historical Evolution of Call Money (daily) and 28-day T-bill Rates: Jan 1997 to Jun 2007
40
38
)
36
Call rate
28-day t-bill rate

(
%
34
32
a
te
30
28
-
b
i
ll R
26
24
22
20
2
8
-
d
a
y

T
18
n
d
16
y a
14
12
one
10
8
a
ll M
6
C
4
2
0
9
7
8
9
9
1
3
6
00
0
1
0
2
0
4
05
0
6
n
-
9
7
n
-
9
7
v-
p
-
9
8
b
-
9
9
l
-
9
9
t
-
00
n
-
0
2
n
-
0
2
v-
p
-
0
3
b
-
0
4
l
-
0
4
t
-
05
n
-
0
7
n
-
0
7
Ja
ec-
a
r-0
ec-
a
r-0
Ju
p
r-9
Ju
a
y-
ug-
p
r-0
Ju
a
y-
ug-
No
A
Se
Fe
D
Ja
Ju
Ja
Ju
M
Oc
M
A
No
A
Se
Fe
D
M
Oc
M
A

Source: Bangladesh Bank
Month-by-month analysis for the period 1997 to 2007 suggests that call rates have an upward
trend starting from March to June and the standard deviations are high in these months
(Table 1). Thereafter, a gradual downward movement sets in up to November. After a brief
slowdown, the rates again move up for the month in December for the practice of ‘window
dressing’ by the commercial banks. From January to March, the rates gently firm up again.
As compared with the 28-day T-bill rate, it is apparent from Figure 1 that call money rate
was lower than the policy rate on different occasions, especially in the earlier periods.
However, this is the daily call rate, which went below the central bank policy rate in different
individual days; the monthly average call rate always remains above the 28-day T-bill rate in
Bangladesh (Table 1).
One important fact regarding the pattern of call money rate is that the variance does not
remain constant over the periods. As can be seen from Figure 2, volatility in the call money
rate follows clustering at some particular points of time; there are periods of high volatility
followed by periods of low volatility. In recent years, volatility in the call rate saw a break
during Feb-Jun 2003. This is the period when the floating exchange rate system was
introduced in Bangladesh.8 In the period immediately before the change, BB tightened
liquidity in the market as a precautionary measure to curb speculative tendencies in the

8 Bangladesh stepped into fully market based exchange rate effective from 31 May 2003.

9

---