Anda di halaman 1dari 15

Asset and liability management in

nancial crisis
Arzu Tektas, E. Nur Ozkan-Gunay and Gokhan Gunay
Department of International Trade, School of Applied Disciplines,
Bog azici University, Istanbul, Turkey
Abstract
Purpose An efcient asset-liability management requires maximizing banks prot as well as
controlling and lowering various risks. This multi-objective decision problem aims to reach goals such
as maximization of liquidity, revenue, capital adequacy, and market share subject to nancial, legal
requirements and institutional policies. This paper models asset and liability management (ALM) in
order to show how different managerial strategies affect the nancial wellbeing of banks during crisis.
Design/methodology/approach A goal programming model is developed and applied to two
medium-scale Turkish commercial banks with distinct risk-taking behavior. This article brings new
evidence on the performance of emerging market banks with different managerial philosophies by
comparing asset-liability management in crisis.
Findings The study has shown how shifts in market perceptions can create trouble during crisis,
even if objective conditions have not changed.
Originality/value The proposed model can provide optimal forecasts of asset-liability components
and banks nancial standing for different risk-taking strategies under various economic scenarios.
This may facilitate the preparation of contingency plans and create a competitive advantage for bank
decision makers.
Keywords Assets management, Liability, Financial management, Banking, Turkey
Paper type Case study
Introduction
Internationalization and nancial integration and have increased the possibility of
nancial contagion among emerging and developed countries. The recent nancial
crises in Asia, Russia, Latin America, and Turkey demonstrated the extent of
vulnerability in global nancial markets to changes in emerging nancial markets.
The signicance of the banking sector in the smooth and efcient functioning of the
overall economy as well as the domestic nancial system has become even more
apparent during these crises. Therefore, developing appropriate bank management
strategies seems crucial for lowering the devastating effects of crises.
A vital issue in strategic bank planning is asset and liability management (ALM),
which is the assessment and management of endogenous nancial, operational,
business and exogenous risks. The objective of ALM is to maximize returns through
efcient fund allocation given an acceptable risk structure. ALM is a multidimensional
process, requiring simultaneous interactions among different dimensions. If the
simultaneous nature of ALM is discarded then decreasing risk in one dimension may
result in unexpected increases in other risks.
ALM has changed signicantly in the past two decades with the growth and
integration of nancial institutions and the emergence of new nancial products and
services. New information-based activities and nancial innovation increased types of
The Emerald Research Register for this journal is available at The current issue and full text archive of this journal is available at
www.emeraldinsight.com/researchregister www.emeraldinsight.com/1526-5943.htm
This article is supported by Bogazici University Research Fund.
ALM in
nancial crisis
135
The Journal of Risk Finance
Vol. 6 No. 2, 2005
pp. 135-149
qEmerald Group Publishing Limited
1526-5943
DOI 10.1108/15265940510585806
endogenous and exogenous risks as well as the correlation between these.
Consequently, the structure of balance sheet instruments has become more complex
and the volatility in the banking systemhas increased. These developments necessitate
the use of quantitative skills to manage risks more objectively and improve
performance:
Banking risk management is both a philosophical and operational issue. As a philosophical
issue, banking risk management is about attitudes toward risk and the pay off associated
with it, and strategies in dealing with them. As an operational issue, risk management is
about the identication and classication of banking risks, and methods and procedures to
measure, monitor and control them (Angelopoulos and Mourdoukoutas, 2001, p. 11).
Diversity in bank decision makers attitudes toward risk results in diverse ALM
strategies. Risk taker decision makers are willing to accept higher risk for higher
return whereas risk averse managers accept lower level of risks for lower return.
Consequences of high risk taking strategies might be more devastating in unstable
macroeconomic environments such as emerging nancial markets, leading to systemic
banking crises. On the other hand, nancial risks may also increase a banks overall
risk. Since this type of interdependency has been observed during the recent crises in
Turkey, it might be worthwhile to analyze ALM strategies with varying risk-taking
attitudes during the crises period in order to demonstrate the sensitivity of banks
performance to different risk taking strategies.
The aim of the paper is to model ALM in order to show how different managerial
strategies affect the nancial well being of banks during crisis. The goal programming
model is built on the managerial goals and policies of the bank along with nancial and
regulatory conditions. It manages different risks such as liquidity, capital, credit,
currency, market and operational risks that comprise critical issues in emerging
nancial markets. Banks carrying these risks and facing worsening economic
conditions could benet using the proposed model by analyzing the consequences of
alternative options in asset-liability management. The model is applied to two medium
scale Turkish commercial banks, one risk-averse and the other risk-taker, for the crisis
year 2000 as a specic case of an emerging nancial market.
The nancial structure of the Turkish banking sector during crisis is discussed in
the next section. Banking literature on multi-objective decision making; methodology
and the model developed; empirical results and conclusion are discussed in consecutive
sections.
The Turkish banking sector in the crises years
The Turkish banking sector has faced challenges calling for restructuring with new
rules and regulations since 1980. During this period, Turkey experienced a number of
nancial crises that severely hit the economy and the banking sector. Among these
crises, the recent twin crisis of November 2000 and February 2001 are the most severe
ones leading to a systemic banking crisis, in combination with a currency and an
economic crisis. Twin crises were due to a decade long macroeconomic problems such
as high ination, high public sector borrowing, credit rationing in the private sector,
rapid growth of the banking sector without prudential regulatory institutions, foreign
exchange open positions in the banking system, and dominance of inefcient large
state banks (Ozkan-Gunay, 1998). At the end of 1999, the government announced an
economic program, Disination Program, based on a pre-determined exchange rate
path as a nominal anchor, tight scal and monetary policies, and important structural
JRF
6,2
136
reforms that were postponed for years. All economic agents demonstrated an
optimistic approach and the public fully supported the announced program. However,
the Economic Program failed in the last quarter of 2000 due to the political conicts
prevailing in the coalition government and the delay in structural reforms. Foreign
investors transferred their funds abroad. The high foreign exchange demand and the
control of liquidity by the Central Bank led to the collapse of the payment system in the
banking sector. Especially the crash in large state banks had a chain effect in the
overall economy as well as in the banking sector. Public loss of trust resulted in sharp
declines in bond and stock prices and withdrawals from banks, triggering the twin
crises. The impacts of the twin crises were severe enough to justify government take
over, closing or partial intervention in 23 banks as of June 2002. In the real sector, many
small size enterprises went bankrupt and many large rms faced liquidity problems.
Under the above circumstances, the structural composition of assets and liabilities
changed in the balance sheets of the banks over the years. FX deposits became popular
because economic agents expected that the pre-determined exchange rate path and the
real appreciation of the Turkish Lira (TL) would lower costs of borrowing in FX
liabilities. On the other hand, the high public sector borrowing led to high interest rates
on government securities making these instruments more attractive for the banking
sector. Banks were carrying open FX positions because high domestic real interest
rates made it attractive to borrow abroad and buy government securities to nance
public sector decit. The share of government securities in total assets increased from
10 percent in 1990 to 23 percent in 1999 (The Banking Regulation and Supervision
Agency, 2001). The real sector was crowded-out. The credit to gross national product
(GNP) ratio was about 25 percent, the lowest among all emerging market economies
(The Banking Regulation and Supervision Agency, 2001). As a result, the share of the
liquid assets declined while the share of government securities and loans, especially
consumer credits increased in bank balance sheets. A mismatch of maturities resulted
from an increase in short-term bank funding sources in liabilities as well as an increase
in the percentage of FX liabilities in total liabilities.
The unexpected factors affected the sectors as well as the banks nancial
performance. Deteriorating nancial ratios in the banking sector make the impact of
the crisis apparent (Table I). An important characteristic of 2000 is banks preference to
carry a high ratio of government securities in their portfolio. Although the securities
have high returns, they can be considered as risky investment due to liquidation
problems. It would be difcult to resell these securities in the secondary market due to
high loss especially during crisis. Decline in liquidity and government securities
portfolio gures is a consequence of this development. The decline in return on asset
(ROA) is the result of the imposition of a capital tax on government securities, the
increase in provisions due to increase in non-performing loans and the additional tax
because of the earthquake.
The vulnerability of the banking sector to economic distortions makes it worthwhile
to analyze the situation at individual bank level especially for banks with diverse
policies. Two medium scale banks operating in the Turkish banking sector are chosen
as representatives of distinct risk taking behaviors. These banks are referred to as risk
averse bank (RAB) and risk taker bank (RTB). Table I shows the nancial standing of
RAB and RTB compared to sub groups of banks as well as the sector during the crisis.
RTB prefers to borrow short-term FX funds and lend in longer-term TL assets or buy
government securities. This creates a maturity mismatch and a high foreign exchange
open position in its balance sheet. RAB chooses to maintain high levels of liquidity and
ALM in
nancial crisis
137
Y
e
a
r
s
S
t
a
t
e
P
r
i
v
a
t
e
F
o
r
e
i
g
n
R
A
B
R
T
B
S
e
c
t
o
r
C
a
p
i
t
a
l
a
d
e
q
u
a
c
y
r
a
t
i
o
(
S
h
a
r
e
h
o
l
d
e
r
s

e
q
u
i
t
y
/
T
o
t
a
l
r
i
s
k
-
w
e
i
g
h
t
e
d
a
s
s
e
t
s
)
1
9
9
9
2
0
0
0
1
1
.
7
7
.
8
1
7
.
2
1
8
.
4
2
2
.
5
1
7
.
8
1
0
.
1
1
1
.
2
2
2
.
4
6
.
4
8
.
2
1
3
.
4
L
i
q
u
i
d
i
t
y
r
a
t
i
o
(
L
i
q
u
i
d
a
s
s
e
t
s
/
T
o
t
a
l
a
s
s
e
t
s
)
1
9
9
9
2
0
0
0
2
3
.
2
1
8
.
3
4
3
.
5
3
6
.
5
6
3
.
6
5
9
.
1
5
9
.
7
5
0
.
6
3
9
.
8
3
5
.
8
3
5
.
9
3
2
.
1
C
r
e
d
i
t
r
i
s
k
(
N
o
n
-
p
e
r
f
o
r
m
i
n
g
l
o
a
n
s
/
T
o
t
a
l
l
o
a
n
s
)
1
9
9
9
2
0
0
0
1
0
.
0
1
2
.
5
3
.
6
6
.
1
2
.
7
2
.
9
0
.
5
0
.
9
1
.
5
5
.
5
1
0
.
7
1
1
.
5
C
u
r
r
e
n
c
y
r
i
s
k
(
F
X
a
s
s
e
t
s
/
F
X
l
i
a
b
i
l
i
t
i
e
s
)
1
9
9
9
2
0
0
0
9
9
.
2
9
3
.
8
8
2
.
2
7
4
.
2
7
5
.
4
7
2
.
6
8
7
.
2
8
7
.
4
6
8
.
6
8
1
.
0
7
9
.
4
7
6
.
0
R
e
t
u
r
n
o
n
a
s
s
e
t
s
(
N
e
t
i
n
c
o
m
e
/
A
v
e
r
a
g
e
t
o
t
a
l
a
s
s
e
t
s
)
1
9
9
9
2
0
0
0
1
.
5
1
.
2
5
.
6
0
.
7
8
.
2
0
.
7
3
.
9
2
.
8
8
.
8
2
5
.
6
2
0
.
6
2
3
.
1
A
s
s
e
t
s
q
u
a
l
i
t
y
(
T
o
t
a
l
l
o
a
n
s
/
T
o
t
a
l
a
s
s
e
t
s
)
1
9
9
9
2
0
0
0
2
4
.
3
2
5
.
8
3
3
.
5
3
7
.
7
1
6
.
5
1
7
.
1
1
7
.
6
2
2
.
6
2
7
.
0
4
9
.
2
3
0
.
1
3
2
.
8
I
n
t
e
r
e
s
t
i
n
c
o
m
e
/
I
n
t
e
r
e
s
t
e
x
p
e
n
s
e
1
9
9
9
2
0
0
0
1
0
9
.
8
1
1
4
.
0
1
7
1
.
6
1
7
7
.
5
1
7
5
.
8
1
5
8
.
1
2
2
4
.
7
2
1
9
.
2
1
7
5
.
1
1
0
6
.
3
1
2
9
.
5
1
2
9
.
6
S
h
a
r
e
o
f
g
o
v
e
r
n
m
e
n
t
s
e
c
u
r
i
t
i
e
s
i
n
t
o
t
a
l
a
s
s
e
t
s
1
9
9
9
2
0
0
0
1
0
.
7
6
.
2
2
0
.
9
1
0
.
5
3
1
.
5
1
2
.
9
2
1
.
7
1
7
.
9
2
6
.
5
3
0
.
8
1
7
.
2
1
1
.
5
M
a
r
k
e
t
s
h
a
r
e
i
n
d
e
p
o
s
i
t
s
1
9
9
9
2
0
0
0
3
9
.
8
4
0
.
3
4
6
.
4
4
3
.
5
2
.
7
3
.
2
0
.
6
0
.
6
2
.
3
2
.
5
M
a
r
k
e
t
s
h
a
r
e
i
n
l
o
a
n
s
1
9
9
9
2
0
0
0
2
8
.
2
2
7
.
0
5
5
.
1
5
4
.
5
2
.
9
2
.
8
0
.
5
0
.
7
2
.
0
3
.
6
M
a
r
k
e
t
s
h
a
r
e
i
n
t
o
t
a
l
a
s
s
e
t
s
1
9
9
9
2
0
0
0
3
4
.
9
3
4
.
2
4
9
.
5
4
7
.
4
3
.
5
3
.
6
0
.
9
1
2
.
2
2
.
4
S
o
u
r
c
e
:
B
a
n
k
s
i
n
T
u
r
k
e
y
,
1
9
9
9
a
n
d
2
0
0
0
Table I.
Ratio analysis of the
banking sector (%)
JRF
6,2
138
capital adequacy while giving high priority to the quality of their loan portfolios and
taking low currency risks. Financial performance of two banks relative to different
bank groups is summarized for pre-crisis and crisis year in Table I.
These gures indicate that RAB performed above sector average in most signicant
performance measures such as capital adequacy, liquidity, credit risk, currency risk,
and return on assets in pre-crisis period. RAB did not experience any deterioration in
any of the measures despite the worsening ratios of the sector in the crisis year. Capital
adequacy of RAB improved from 10 percent to 11 percent and there was only a
minimal increase in its credit risk ratio from 0.5 percent to 0.9 percent compared to 11.5
percent risk ratio of the sector. On the other hand, RTB, which performed similar to
RAB before the crisis, experienced a sharp deterioration in all its measures. Capital
adequacy ratio of RTB decreased from 22 percent to 6 percent in the crisis year and the
credit risk ratio increased from1.5 percent to 5.5 percent for the same bank. ROA ratios
of RAB showed little deviation with 4 percent and 3 percent for the years 1999 and
2000 compared to the respective values of 9 percent and 26 percent for RTB. RABs
government securities portfolio is approximately equal to the private banks average
level for both years; on the other hand, RTBs securities portfolio is above the private
sector in 1999 and almost triple in 2000. Neither RAB nor RTB market shares in
deposits, loans, and total assets show any signicant difference before and during
crisis.
Since the vulnerability of the Turkish economy affects the banking sector deeply,
bank managers in Turkey should consider various scenarios in preparing their
proforma budgets. In this paper, a multi-objective decision model is suggested as a tool
to that end. The proposed model can provide optimal forecasts of asset-liability
components and banks nancial standing for different risk taking strategies under
various economic scenarios. This may facilitate the preparation of contingency plans
and create a competitive advantage for bank decision makers.
Banking literature on multi-objective decision making
The literature has recorded a number of mathematical programming models to
coordinate asset-liability planning using optimal decision strategies. The models have
been single or multiple goal, deterministic or stochastic. Deterministic models have
found more acceptability because of practitioners unfamiliarity and computational
difculties related with stochastic models. The operational models assume that a bank
seeks to maximize its expected prots subject to portfolio mix constraints. Myers
(1968) and Hester and Pierce (1975) conclude that the appropriate objective function for
a nancial institution involves maximization of the expected net present value.
The deterministic models follow the linear programming model approach of
Chambers and Charnes (1961). Their study maximizes net discounted returns, subject
to budget, liquidity and system constraints. Further applications of their model are
Cohen and Hammer (1967), Komar (1971). Single objective nature of these models
ignores the multiple objectives in banking decisions. Zopounidis (1999) justies the
multi-criteria character of nancial problems such as optimization of assets and
liabilities.
Goal programming and other multi-objective decision models have been applied to
banking by Forston and Dince (1977), Booth and Dash (1979), Eatman and Sealey
(1979), Korhonen (1987), Booth and Bessler (1988). Booth and Dash (1979), Booth and
Bessler (1988) and Korhonen (1987) address the multistage nature of the problem;
ALM in
nancial crisis
139
Booth and Bessler (1988) take into account interest-rate risk as well. Forston and Dince
(1977) use a goal programming approach where four competing goals (prot,
loan-deposit ratio, capital adequacy, liquidity) are considered. Eatman and Sealey
(1979) consider three objectives, namely net prot maximization, capital-adequacy
ratio and risky assets-capital ratio minimization. Korhonen (1987) applies a two-stage
goal programming model to the management of domestic and foreign currency
dominated assets and liabilities of a bank. The goals deal with expected prots, risk,
liquidity, capital adequacy, growth and other aspects of the banks operations. Giokas
and Vassiloglou (1991) discuss the construction and application of a goal programming
model that takes into account the essential institutional, nancial, legal and bank
policy considerations.
Among major studies on bank asset-liability management in the Turkish banking
sector, Oguzsoy and Guven (1997) develop a stochastic linear model, which determines
the portfolio of assets and liabilities given a set of deterministic rate of returns and
costs of borrowings and a set of random deposit levels, liquidity, and total reserve
requirements. Guven and Persentili (1997) develop a linear programming model, which
aims at determining an optimal sequence of balance sheet positions for a bank. Both
models are implemented using Turkish banking sector data for 1987-1990 period.
Methodology
There is no way simultaneously to maximize returns (or prots) and minimize risks
but banks can only make risk/return tradeoffs and attempt to maximize returns for
whatever aggregate level of risk they choose to undertake (Uyemura and Van
Deventer, 1993). The objective of goal programming is to reach a satisfactory level of
multiple objectives whenever it is not possible to achieve every goal to the full extent,
so that the decision maker may come as close as possible to reaching goals.
Goal programming model (GP) is preferred to nd the optimal composition of a
banks assets and liabilities in the Turkish banking sector. Choice of GP best describes
the multi-objective nature of the problem and it eliminates the computational
difculties as well as practitioners unfamiliarity experienced with stochastic models.
Flexible nature of GP enables the decision maker to incorporate a number of goals
under a set of constraints. GP minimizes deviations between set goals and what can
actually be achieved within the given constraints.
In GP, the objective function (1) minimizes the sum of deviations (d
i
) from each goal
(G
i
). Each goal is assigned a priority weight (P
i
) that shows its relative importance
among other goals. Therefore, goals with higher P
i
values are achieved before the
others. The goals (2) reect the objectives that are set by decision makers. Constraints
(3) represent the availability or upper/lower limits of resources.
Determine X x
1
; x
2
; . . .; x
j
; . . .x
n
such that:
Min Z f d

i
; d
2
i


X
I
i1
P
i
d

i
; d
2
i

1
subject to:
Goals G
i

X
n
j1
a
ij
x
j
d
2
i
2d

i
b
i
for i 1; :::; I : 2
JRF
6,2
140
Constraints
X
n
j1
c
mj
x
j
# r
m
for m 1; :::; M 3
where:
x
j
: mean value of variable j;
a
ij
: technological coefcient of x
j
in goal i;
b
i
: target value of goal i;
c
mj
: consumption coefcient of x
j
in constraint m;
r
m
: available amount of resource m;
d

i
: overachievement of the target for goal i;
d
2
i
: underachievement of the target for goal i; and
P
i
: priority weight of the deviation variables of goal i.
Asset-liability management model (ALMM)
GP is applied to the Turkish banking sector taking into account sector specic
characteristics. The ultimate aim is to identify the best possible composition of a
banks assets and liabilities by controlling the various types of risks. The model
contributes to similar studies in literature by including emerging nancial market
specic goals and constraints as well as testing the sensitivity of bank performances
for different risk taking strategies during crisis.
ALMM is developed as follows:
Min Z P
1
d
2
1
P
2
d
2
2
P
3
d
2
3
P
4TL
d
2
4TL
P
4FX
d
2
4FX
P
5TL
d
2
5TL
P
5FX
d
2
5FX
P
6
d
2
6
subject to:
(1) Goals:
.
G1: Liquidity (Liquid assets/Total assets) d
2
1
2d

1
a.
.
G2: Revenue (Total interest income 2 Total interest expense)
d
2
2
2d

2
b.
.
G3: Capital adequacy (Shareholders equity/Total risk weighted assets)
d
2
3
2d

3
c.
.
G4
TL
: Deposit
TL
market share (Total TL deposits
bank i
/Total
TL deposits
sector
)d
2
4TL
2d

4TL
d.
.
G4
FX
: Deposit
FX
market share (Total FX deposits
bank i
/Total FX
eposits
sector
) d
2
4FX
2d

4FX
e.
.
G5
TL
: Loan
TL
market share (Total TL loan
bank i
/Total TL loan
sector
)
d
2
5TL
2d

5TL
f .
.
G5
FX
: Loan
FX
market share (Total FX loan
bank i
/Total FX loan
sector
)
d
2
5FX
2d

5FX
g.
.
G6: Total asset Total assets d
2
6
2d

6
h.
ALM in
nancial crisis
141
(2) Environmental constraints:
.
E1: Capital Adequacy ratio $ 8 percent.
.
E2: TL reserve requirement/TL total deposits $ 6 percent.
.
E3: FX reserve requirement / FX total deposits $ 11 percent.
.
E4: Disponibility for TL $ 2 percent.
.
E5: Disponibility for FX $ 2 percent.
.
E6: Disponibility for government bonds as legal reserves $ 4 percent TL
deposits 1 percent FX deposits.
.
E7: Shareholders equity/Total liabilities $ 5 percent.
.
E8: Total loans/Total assets $ 30 percent.
(3) Bank policy constraints:
.
P1: Market risk a # Securities=Total assets # b.
.
P2: Currency risk t # FX assets=FX liabilities # d:
.
P3: Currency risk 1 # TL assets=Total assets # f:
.
P4: Credit risk g # Non 2performing loans=Total loans # h.
.
P5: Interest rate risk Short termloans=Total loans $ w.
.
P6: Liquidity risk TL cash=Total TL assets $ l.
.
P7: Liquidity risk FX cash=Total FX assets $ u:
.
P8: Liquidity risk TL due from banks=Total TL assets $ r.
.
P9: Liquidity risk FX due from banks=Total FX assets $ v.
.
P10: Liquidity risk Liquid assets=Total deposits $ m.
.
P11: Liquidity risk Current assets 2 Current liabilities/Total assets $ p.
.
P12: Operational risk Fixed assets=Total assets # s:
.
P13: Operational risk Operating expenses=Total assets # c:
(4) Technical constraint. T1: Total assets Total liabilities.
The ratios dened in goals and constraints are operationalized using decision variables
that are generally accepted balance sheet items of a Turkish bank (Table II). As an
example, G1 is operationalized as:
X1
TL
X1
FX
X2
TL
X2
FX
X3
TL
X3
FX
::: X11
TL
X11
FX
Y1
TL
Y1
FX
Y2
TL
Y2
FX
Y3
TL
Y3
FX
::: Y11
TL
Y11
FX
$ a
where a is set as 0.40 by RAB and 0.30 by RTB.
Goals (G1-G6) are dened as meeting minimum levels of liquidity, revenue,
capital adequacy as well as market shares in total assets, deposit and loan, both TL
and FX. Right-hand-side values of goals are determined based on the related banks
policies.
Constraints are classied as environmental and bank policy constraints.
Environmental constraints include the regulatory constraints imposed by the
monetary authorities and the market driven constraints. Based on Basel risk-based
JRF
6,2
142
capital criteria, banks should maintain a minimum capital adequacy level of 8 percent
against losses, which may result from existing and potential risks (R1). The Central
Bank of Turkey (The Central Bank of Turkey, 1996) requires a minimum TL and FX
reserve requirements of 4 percent (E2) and 11 percent (E3), respectively. Minimum TL
and FX disponibilities (required reserves for liability items excluding deposits and own
Decision variables
Risk-averse bank Risk-taker bank
OS PS Actual OS PS Actual
X1
TL
Cash TL 2.6 2.4 3 7 8.2 16
X1
FX
Cash FX 34.4 27.7 40 15.6 50.2 27
X2
TL
Due from banks TL 26.6 23.6 73 47.2 62.5 37
X2
FX
Due from banks FX 172.1 138.3 316 46.7 104 77
X3
TL
Central bank TL 2.6 2.2 2 12 6.8 4
X3
FX
Central bank FX 15 12.3 3 40 38.5 10
X4
TL
Other nancial institutions TL 0 0 0 0
X4
FX
Other nancial institutions FX 0 0 0 0
X5
TL
Interbank funds sold TL 17.6 14.6 60 574.6 193.1 16
X5
FX
Interbank funds sold FX 525.5 354 207 205.6 680.1 0
X6
TL
Government securities TL 237.1 237 1 827 827 783
X6
FX
Government securities FX 121.1 121 5 368 368 289
X8 Other securities 0 0 0 0 0 77
X9
TL
Reserve requirements TL 7.9 6.5 3 36 20.4 0
X9
FX
Reserve requirements FX 82.5 67.7 60 220 201.7 1
X10
TL
Short-term loans TL 352.9 247.9 100 668.4 608.4 741
X10
FX
Short-term loans FX 190.9 197.6 173 507.1 443.8 788
X11
TL
Long-term loans TL 0 43.3 53 0 0 288
X11
FX
Long-term loans FX 0 0 20 0 0 17
X12 Non-performing loans 11.1 6.2 3 47.5 47.5 100
X13 Equity participations (nancial) 0 104 1 84 84 164
X14 Other participations 0 0 0 0 0 10
X15 Afliated companies 0 0 298 66 66 168
X16 Fixed assets 0 0 13 100 100 140
X17
Government bonds account funding
legal reserves 12.8 10.5 0 32.8 0 44
X18 Accrued income 37 33 83 80 89.8 59
Y1
TL
Demand deposit TL 88.2 72.8 26 130 226.4 35
Y1
FX
Demand deposit FX 132.4 39.5 70 180 339.6 106
Y2
TL
Time deposit TL 44.1 36.2 89 470 113.2 1415
Y2
FX
Time deposit FX 617.6 576.4 470 1820 1584.7 995
Y3
TL
Interbank funds TL 0 0 1 0 0 45
Y3
FX
Interbank funds FX 0 105 0 0 0 0
Y4
TL
Central bank funds borrowed TL 0 0 0 0
Y4
FX
Central bank funds borrowed FX 0 150 0 0 0 0
Y5 Domestic bank funds borrowed 0 0 21 0 0 33
Y6 Abroad bank funds borrowed 518.5 150 424 659.6 686.7 578
Y7 Other funds borrowed 200.1 289 158 0 456 27
Y8 Accrued interest 56 45.2 67 130.4 122.3 75
Y9 Premium payable 4 4 4 48 48 70
Y10 Provision 15 15 15 144 144 31
Y11 Shareholders equity 100 100 103 258 186 461
Y12 Prot/loss 74 66 35 160 93 Loss
Table II.
Empirical results under
different scenarios
ALM in
nancial crisis
143
funds) of 2 percent each (E4, E5) should be met. Disponibility for government bonds
should constitute at least the sum of 4 percent of TL and 1 percent of FX deposits (E6).
Finally, a minimum 5 percent ratio of shareholders equity to total liabilities (E7) and a
minimum 30 percent of total loans to total assets ratio (E8) are the generally accepted
principles in the Turkish banking sector.
Bank policy constraints capture the major risks present in emerging nancial
markets. Their upper and lower limits reect bank management strategies and can be
adjusted to changing preferences and environmental conditions. They are classied
information therefore they are not presented in the paper. Market risk includes external
factors such as environmental, political and monetary factors. Since high public sector
borrowing requirement has been an important constraint for the government, the
model controls market risk stemming from borrowing requirement by assigning upper
and lower limits on the government securities held by the bank (P1). Currency risk
shows the arbitrage between FX borrowing funds and TL asset items and is controlled
by imposing a range for the ratio of assets to liabilities in FX and TL (P2 and P3).
Credit risk represents the possibility of a delay in a repayment or no repayment of
credit. The constraint (P4) limits the ratio of non-performing loans to total loans.
Interest rate risk is the mismatch of maturities between short-term borrowings and
long-term lending as well as the pricing mismatch of the assets and liabilities in terms
of xed and variable interest rate. The related constraint (P5) imposes a minimum level
on the amount of short-term loans. Liquidity shows how quickly assets can be
converted to cash to meet current obligations. Different types of liquid items such as
cash and due from banks are forced to maintain a minimum level (P6-P11). Operational
risk is controlled by maximum upper limits of xed assets and operating expenses in
total assets (P12 and P13).
To demonstrate the impact of different risk taking behaviors on bank
performance during crisis, ALMM is run for two commercial banks with opposite
strategies. Risk behavior differences are reected in the model by changing the
target levels of related goals and upper and lower limits of bank policy constraints
as shown in Table III.
Liquidity and capital adequacy goals are taken as indicators of attitudes towards
risk. Minimumcompulsory risk levels for liquidity and capital adequacy are dened by
the monetary authorities. Although RTB sufces with these levels, RAB prefers to
keep higher ratios in order to lower its risk under unstable economic conditions. RTB
takes higher market risk by carrying higher volume of government securities in total
assets. On the other hand, RAB lowers currency risk by holding higher ratio of FX
assets. RAB is more sensitive to remain liquid therefore; it aims at higher ratio of liquid
assets in total assets.
G1 (%)
liquidity
goal
G3 (%)
capital adeq.
goal
P1 (%)
market
risk
P2 (%)
currency
risk
P3 (%)
currency
risk
P10 (%)
liquidity
risk
P11 (%)
liquidity
risk
RAB 40 10 15-20 80-90 25-35 70 8
RTB 30 8 25-35 45-60 65-75 25 5
Table III.
Risk behaviors of the
representative banks
JRF
6,2
144
Data and empirical results
Financial data for 1999 and 2000 were obtained from the annual publications of
Banking Association of Turkey. The data included the balance sheets and income
statements of the two banks analyzed in this paper. Face-to-face, semi-structured
interviews were carried with the vice-presidents of banks to collect data about bank
goals and strategies. Given the condential nature of the required data, existing
network of contacts was used in order to reduce initial problems of access. They were
requested to provide information on the monetary and non-monetary targets of their
banks for pre-crisis and crisis periods. Information provided was used to structure
goals and bank policy constraints of each bank.
The year 2000 can be characterized as a break point in terms of realization of
expectations in the sector. It started with optimistic expectations based on the
Economic Program, but ended with unexpected crises due to failures in the
implementation of the announced program. Therefore, it is important to see the impact
of nancial decisions made in the optimistic environment of the early 2000 on the
wellbeing of the banks under the worsening end of the year conditions. Therefore,
ALMM is run for two banks, RAB and RTB, for two different economic scenarios for
the crisis year 2000. First scenario is an optimistic scenario (OS), which assumes the
success of the Economic Program whereas the second one is a pessimistic scenario
(PS), assuming the failure of the Program. There are some major differences between
OS and PS versions of ALMM. In PS, the banks are expected to adjust their balance
sheet items accordingly. Revenue goal is affected because interest income and
expenditures change due to higher interest rates under deteriorating market
conditions. Economic agents prefer to remain more liquid under such conditions.
Besides, it is also assumed that the bank would not be able to resell the government
securities due to the high loss and/or the problems in the secondary market. Therefore,
lower limits of liquidity related constraints are reduced. An upper limit is imposed on
the amount of foreign borrowing due to the high premium risk at international
nancial markets.
ALMM is run for RAB and RTB under both scenarios by assigning equal priority
weights for each goal. Results for OS, PS and the actual balance sheets of 2000 are
presented in Table II. Model results are discussed for RAB and then for RTB. Table IV
presents actual goals set by bank managers and model results dened as deviations
from the goals for RAB.
Two of the overachieved goals are liquidity and capital adequacy, which are
important indicators of risk-averse behavior. Liquidity is 5 percent; capital adequacy
ratio is 5 percent above the target and 7 percent above the Basel criteria. The market
share in TL deposits is also over-achieved by 6 percent and the market share in TL
loans is 18 percent above the target. Market share target of FXdeposits and total assets
G1 G2 G3 G4
TL
G4
FX
G5
TL
G5
FX
G6
Goal 40% 136 10% 125 750 305 500 1,800
d

8% 0 5% 7.5 0 55 0 0
d
2
0 19 0 0 0 0 230 0
Note: Values other than G1 and G3 are in USD million
Table IV.
Goal achievement levels
for the risk-averse bank
ALM in
nancial crisis
145
are just achieved. Interest revenue goal is under-achieved by 14 percent, the market
share in FX loans is 46 percent below the target.
Some major results of the bank policy constraints also strengthen the risk-averse
behavior of the bank. Liquidity risk measure (P10) is as high as 90 percent. This
implies that bank prefers to remain liquid while preferring high return liquid assets
such as lending FX in interbank and having FX accounts at other banks. Credit risk
measure (P4) is also minimized allowing at most 2 percent non-performing loans in
total loans. Measure of interest rate risk (P5) is as high as 98 percent. This behavior
shows that the bank is sensitive to maturity mismatch. Currency risk measure (P2) is
90 percent, implying that the bank is minimizing currency risk by avoiding the foreign
exchange open position. Results of OS show that 85 percent (USD750 million) of the
liabilities are collected in FX while 62 percent (USD1.148 million) of assets are kept as
FX. Market risk measure (P1) is low with a ratio of 19 percent government securities in
total assets. On the other hand, the model proposes to invest in TL government
securities due to the high returns and offers TL loans through open position. The
remaining FXliabilities are allocated in FX interbank funds and FX short-termloans in
the model. OS results in 4 percent ROA compared to 2.3 percent in the actual balance
sheet. A positive ROA is striking for the year 2000 compared to a sector average of
23.1 percent.
PS results for RAB show deteriorations compared to OS results. Liquidity ratio
drops from 48 percent to 39 percent, revenue decreases by 19 percent, total assets
shrink by 11 percent. Market share in TL deposits and TL loans, which are
overachieved in Run 1 are underachieved by 10 percent and 3 percent, respectively.
Since it is difcult to borrow abroad due to the country risk, the bank can use this
resource in a limited amount (USD150 million). Another difference is a 32 percent
decline of FX interbank funds sold along with an increase in long-term TL loans. There
is no considerable change in the remaining balance sheet items.
Related goal levels and constraint limits for RTB are given in Table III. Table V
presents OS results.
Capital adequacy is the only overachieved goal. Liquidity ratio is 1 percent below
the target. Interest revenue goal is under-achieved by 41 percent, indicating a lower
revenue level (USD145 million) than the target but it is still above the realized revenue
level of RAB. The market share in TL deposits, TL loans, and FX loans are also
under-achieved by 27 percent, 27 percent, 25 percent, respectively. It is difcult to
expand the deposit base in TL where approximately 60 percent of the deposits are held
in FX in the banking system. On the other hand, capital adequacy ratio of 10 percent is
7.5 percent above the target and 9.5 percent above the Basel criteria. The remaining
market share goals meet the target levels.
Considering the major implications of bank policy constraints under OS, RTB takes
high currency and interest rate risk to earn high prots. Currency risk measure (P2) is
G1 G2 G3 G4
TL
G4
FX
G5
TL
G5
FX
G6
Goal 35% 350 10% 825 2,000 914 675 4,000
d

0 0 7.5% 0 0 0 0 0
d
2
1% 145 0 225 0 246 168 0
Note: Values other than G1 and G3 are in USD million
Table V.
Goal achievement levels
for the risk-taker bank
JRF
6,2
146
53 percent, indicating higher open FX position compared to RABs 90 percent. 81
percent (USD2,582 million) of the liabilities are collected in FX while 33.7 percent
(USD1,340 million) of assets are kept as FX. Liquidity measure (P10) of 53 percent is
lower compared to RABs rate of 90 percent. This implies that RTB might take higher
liquidity risk while preferring high return assets such as government securities and
having TL accounts at other banks. This preference is also indicated by a higher
market risk (P1) ratio of 30 percent compared to 19 percent ratio of RAB.
Owing to the expectations in the banking system, longer-term loans are offered
seldom, resulting in maturity mismatch, which also brings the interest rate risk
problem. In order to lower interest rate risk of 96 percent, the model proposes RTB to
allocate its FX liabilities 68 percent in FX time deposits, 25 percent foreign banks
borrowed and 7 percent FX demand deposit. On the asset side, non-performing loans,
which constitute 5.5 percent of total loans in actual (Table I), is 3.8 percent under OS.
When credit risk ratio is compared with the actual ratio of 5.5 percent, the model can
control this risk up to a certain level. OS results in 6 percent ROA but a negative return
of 25.6 percent is realized in the actual balance sheet for the year 2000 compared to a
sector average of 23.1 percent (Table I). This demonstrates that applying the same
strategies by using the proposed model can improve the performance.
Comparing the results of OS and PS for RTB, the main distinction is the decline in
liquidity and capital adequacy goal. Capital adequacy goal declines from 17.3 percent
to 13.5 percent and liquidity goal drops from 35 percent to 33 percent. Since RTB has a
large portfolio of government securities, it cannot sell in the secondary market due to
the liquidity problem in the market. Market share in deposits and loans are
underachieved signicantly. Underachievement of the market share goals in TL
deposit, FX deposits, TL Loan, and FX loan are 58 percent, 4 percent, 33 percent, and
34 percent, respectively. The volume and composition of loans are similar in both
scenarios. The striking difference is that the model proposes to allocate available funds
in FX alternatives, reecting the risk-avoiding behavior in terms of currency and
interest rate risks. It signicantly decreases TL interbank funds borrowed while
tripling FX interbank. This allocation improves currency risk measure from 52 percent
in OS to 72 percent in PS. Total deposits collected decline by 13 percent in PS. Since
there is no change in total liabilities, the decrease in deposits is compensated by
borrowing other funds in addition to the similar amount of foreign funds borrowed in
OS. The prot of RTB is almost halved compared to OS.
Since 2000 is the crisis year, PS results are compared to the actual balance sheet for
two banks. The model can propose rational solutions and improve the nancial well
being of the banks even during crisis. The model suggests RAB to move from a low
risk-low return position to a higher risk-higher return one by increasing the share of
government securities and short-term loans. It also proposes RTB to control currency,
credit, and maturity mismatch risks by lowering the amount of government securities,
increasing the interbank funds sold in FX and lowering the share of non-performing
loans in total loans on the asset side. These changes result in USD93 million prot in
PS compared to a loss in actual for the RTB and an improvement in prot from USD35
million to USD66 million for RAB. On the liability side, RABs actual and proposed
deposit allocations do not exhibit a signicant divergence compared to RTB. The
model emphasizes funding by using abroad banks along with other funding
opportunities. Although RTB requires signicant amount of foreign borrowing or
other sources of borrowing, this would be almost impossible during nancial crisis.
This could be interpreted as the need to attract a new major shareholder so that the
ALM in
nancial crisis
147
risk-taker bank can overcome foreign source deciency and avoid bankruptcy. Owing
to the nancial problems, RTB was taken over by Savings Deposits Insurance Fund at
the end of 2000.
Conclusion
Strong domestic nancial systems play an important role in development and
stabilization of the overall economy. The signicance of the banking sector within the
nancial system becomes even more apparent in emerging markets like Turkey, which
are highly vulnerable to economic distortions and nancial crisis. Structural
differences in emerging markets introduce new risks such as currency and maturity
mismatch thus the bank managers have to consider a wide range of scenarios and
manage their balance sheets optimally by developing an efcient asset-liability
management strategy. This study proposes a multi-objective decision model to reach
an optimal strategy.
GP is developed as a decision tool for ALM in order to show how different
managerial strategies affect the nancial well being of banks during crisis. It is applied
to two medium scale Turkish commercial banks with different risk taking behaviors
and the results are compared under optimistic and pessimistic economic scenarios.
Although applying the model improves wellbeing of both banks, the improvement
becomes more signicant as pessimistic perception of the economic environment or
risk taking propensity of the banks increases. The model is capable of developing
rational responses to strategy and scenario changes, thus can be proposed as a decision
making tool for banks operating under uncertainty. Results could be generalized
further by applying them to representatives of different bank groups as well as testing
the model under various economic and nancial scenarios.
References
Angelopoulos, P. and Mourdoukoutas, P. (2001), Banking Risk Management in a Globalizing
Economy, Greenwood Publishing Group, Westport, CT.
(The) Banking Regulation and Supervision Agency (2001), Towards a Sound Turkish Banking
Sector, The Banking Regulation and Supervision Agency, Ankara.
Booth, G.G. and Bessler, W. (1988), Goal programming models for managing interest-rate risk,
Omega, Vol. 17 No. 1, pp. 81-9.
Booth, G.G. and Dash, G.H. (1979), Alternate programming structures for bank portfolios,
Journal of Banking and Finance, Vol. 3, April, pp. 67-82.
(The) Central Bank of Turkey (1996), Circular on Reserve Requirements and General Disponibility,
No. 96/1, The Central Bank of Turkey, Istanbul.
Chambers, D. and Charnes, A. (1961), Inter-temporal analysis and optimization of bank
portfolios, Management Science, Vol. 7 No. 4, pp. 393-410.
Cohen, K.J. and Hammer, F.S. (1967), Linear programming models for optimal bank asset
management decision, Journal of Finance, Vol. 22 No. 2, pp. 42-61.
Eatman, J.L. and Sealey, C.W. (1979), A multiobjective linear programming model for
commercial bank balance-sheet management, Journal of Bank Research, Vol. 9, Winter,
pp. 227-36.
Forston, J.C. and Dince, R. (1977), An application of goal programming to management of a
country bank, Journal of Bank Research, Vol. 7, pp. 311-19.
JRF
6,2
148
Giokas, D. and Vassiloglou, M. (1991), Agoal programming model for bank assets and liabilities
management, European Journal of Operational Research, Vol. 50 No. 1, pp. 48-60.
Guven, S. and Persentili, E. (1997), A linear programming model for bank balance-sheet
management, Omega, Vol. 25 No. 4, pp. 449-59.
Hester, D.D. and Pierce, J.L. (1975), Bank Management and Portfolio Behavior, Yale University
Press, New Haven, CT.
Komar, R.I. (1971), Developing a liquidity management model, Journal of Bank Research, Vol. 2,
pp. 38-52.
Korhonen, A. (1987), A dynamic bank portfolio planning model with multiple scenarios,
multiple goals and changing priorities, European Journal of Operational Res., Vol. 30,
pp. 13-23.
Myers, S.C. (1968), Procedures for capital budgeting under uncertainty, Industry Management
Review, Vol. 9, Spring, pp. 1-20.
Oguzsoy, C.B. and Guven, S. (1997), Bank asset-liability management under uncertainty,
European Journal of Operational Research, Vol. 102, pp. 575-600.
Ozkan-Gunay, E.N. (1998), Economies of scale and scope in the turkish banking industry:
the effect of the nancial liberalization program, The Journal of Economics, Vol. XXIV
No. 1, pp. 1-18.
Uyemura, D.G. and Van Deventer, D.R. (1993), Financial Risk Management in Banking,
McGraw-Hill, New York, NY.
Zopounidis, C. (1999), Multicriteria decision aid in nancial management, European Journal of
Operational Research, Vol. 119 No. 2, pp. 404-15.
Further reading
(The) Banks Association of Turkey (1998), Banks in Turkey, The Banks Association of Turkey,
Istanbul.
(The) Banks Association of Turkey (1999), Banks in Turkey, The Banks Association of Turkey,
Istanbul.
(The) Banks Association of Turkey (2000), Banks in Turkey, The Banks Association of Turkey,
Istanbul.
ALM in
nancial crisis
149