Transcript The BIS Regulatory Framework and Icelandic Banking Sector

The BIS Regulatory Framework
and Icelandic Banking Sector:
Issues and Dilemmas
Seminar at the Central Bank of Iceland
February 18, 2002
Gudmundur Magnusson
Saso Andonov
1
I. The Original (Old) Basle Accord
(1988) – EU and EEA (1993)
Financial stability
 Levelling of the competitive field
 Minimum requirements
 Standardization
 Credit risk
 Simplicity

2
II. The Basle Proposals for Market
Risk – VAR (1993 and 1995)



Internal Value at Risk models;
Approval by supervisory authorities;
Trade off between CAR and VAR
3
III. The New Basle Accord – Proposals
Still more financial stability;
 More differentiated measures of credit risk
 Other types of risk (liquidity, operational,
legal risk);
 Supervisory review process;
 More transparency;
 Enhanced market discipline

4
IV. Points for Discussion:

Comparison of the New and the Old rules and potential
implications of the new standards for credit risk;
(a) Choice between two alternative approaches: the standardized
approach (SA) and the internal ratings based approach (IRBA).
SA – external credit assessment
IRBA – own internal rating of borrowers (two options: foundation
approach or own estimates)
(b) Supervisors could require higher that the minimum capital target;
(c) IRBA more risk sensitive; good for big banks, bad for small?
5
Price Structure of Credits With and Without
Use of Credit Models
Yield
Practice with credit model
Current practice
Risk
6
Proposals and Practice
Cost of capital versus risk-adjusted rate of
return;
 Credit risk models in practice;
 VaR models, insurance-inspired models,
Option-based models;
 Acceptance by supervisory authorities

7
Points for Disscusion – cont.
(d) More average capital needed as a cushion
for macroeconomic fluctuations?
(e) Small countries and small banks will
suffer?
(f) Manna from heaven for the rating
agencies?
8
V. Comprehensive Risk Management on
the Banking Industry Level








Comprehensive versus narrow scope of banking;
Investment banks and commercial banks;
New types of instruments;
Accounting standards;
Disintegration of banking;
Globalisation;
Lender of last resort;
Regulatory arbitrage
9
D.T. Llewellyn: Assessing the New
Basle Capital Accord

Criteria for judging the effectiveness of any capital
adequacy regime – the following twelve tests are
suggested:
1.
2.
3.
4.
Does it have the effect of aligning the regulatory and
economic capital?
Does it create incentives for effective and efficient risk
analysis, management and control mechanisms;
Does it create appropriate incentives for the correct pricing
of the risk?
Does it creates incentives for an efficient allocation of
capital within the bank?
10
cont.
5.
6.
7.
8.
9.
10.
11.
12.
Does it create perverse incentives for regulatory arbitrage?
Does it create unwarranted entry barriers?
Are the capital requirements competitively neutral?
Are the requirements appropriate for overall portfolio risk,
as opposed to the sum of project risks?
Does it have the effect of impairing competition in banking
markets?
Does it have unfavourable effects on the macroeconomy;
Is the new Accord unnecessarily complicated?
Does it create or reinforce incentives for shareholders and
other official supervisors to monitor the risks of banks and
for market discipline to be exercised?
11
VI. CAR of the Icelandic Banking System:
1989-2001
16
14
12
10
8
6
4
2
0
9
9,7
8,7
9,5
8,9
10,7
13,6
11,1
10,7
9,9
9,8
10,6
9,90
1989
1990 Commercial and Savings Banks
Iceland:
1991
Capital
Adequacy Ratio 1989-2000
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
12
Second: the Alternative One

Financial Supervisory Authority of Iceland:
– "The FSE has also declared that large Icelandic credit institutions, even
showing an effective risk management and internal control, should at least
aim for a minimum of a 10% CAD ratio. Other credit institutions should aim
for a higher CAD ratio". Annual Report, 2000

Central Bank of Iceland:
– "The capital ratio of the financial institutions as a whole has been declining in
recent years. At the end of 2000 it measured 9.9%, down from 10.6% the previous
year. The commercial banks failed to reach their target of maintaining capital
ratios of 10% and above last year". CB of Iceland Monetary Bulletin, Financial
Stability, 2001,

International Monetary Fund:
– "Decline in Icelandic banks’ CARs (both including and excluding subordinated
loans), reflect the continued expansion of banks’ balance sheets without a
corresponding increase in capital, as well as losses sustained on portfolio
13
investments as a result of the emerging pressures in the domestic and foreign
markets". IMF Iceland Financial System Stability Assessment;
Third: From an International
Perspective
Country
Denmark
Sweden
Norway
Iceland
Finland
USA
Japan
Israel
1997
11.61
13.0
12.6
9.8
11.9
12.8
9.1
10.5
1998
11.32
10.8
12.6
9.1
11.5
14
VII. Vulnerability and CAR in the
Nordic Countries

Maintaining credible and prudent levels of CAR may
provide an efficient mechanism in case of worsening
macroeconomic conditions that may drive the system into
banking crises:
– It is a critical indicator of a disturbance of any nature in the
bank operational environment and will ultimately be reflected
in the changes of this ratio;
– It can serve as an early warning signal;


Point for discussion: Would the banking crises in the
Nordic countries in the early 90´s have been less adverse if
agents had maintained higher than the minimum capital
standards?
15
Is the CAR safe at any speed?
For Consideration: CAR in the Early 90’s
Country
Denmark
Norway
Sweden
Finland
Faeroe
Islands
Greenland
Iceland

CAR
Efficiency
10%
Sufficient
< 10%
Insufficient
10%
Insufficient
10%
Sufficient
< 10%
Barely
sufficient
Lesson: Safety of the CAR depends both on its
quality and nature of the environment
16
VIII. The Optimal Size of the CAR


In economic terms, regulatory capital should not be
increased beyond the point where the marginal cost
outweighs the expected marginal benefit from holding
capital;
The most quoted criteria in determining the size of the
CAR are:
– Cost of raising capital;
– Economic cycles and their importance for capital planning;
– Size of the agent and its importance for the volatility of the
income streams;
– Access to liquidity including the Central Bank;
– Credibility and peer group pressures;
17
Modelling the Optimality of CAR



In our opinion, optimality of the CAR should be
derived from basic characteristics of the economy,
both macroeconomic and microeconomic ones;
For that purpose, we are proposing four variables (or
their proxies) to be taken into account when
determining the desired capital ratio above the
regulatory minimum;
The policy rule is based on the weighted differences
or relative differences of the variables from matching
peers (benchmarks);
18
Variable Candidates

Macroeconomic volatility measure:
– Macroeconomic volatility - GDP real growth rate,
price level, productivity, terms of trade;

Diversification measure:
– Sectoral measures - such as loan concentrations
ratios across sectors;
– Overall portfolio measures - such as single exposure
limits;
19
continuing

Microeconomic or banking sector-specific
measure
– Capital strength - such as the size and structure of the
own capital;
– Banking-sector specific volatility, such as bank
deposits, bank credit to the private sector etc;

Profitability measure
– Banking sector efficiency - measured through pre-tax
profits as single measure or through weighted index
of net interest income, non-interest income, operating
expenses and the pre-tax profits.
20
Formally:
CARIC  1 ( IC   AV )   2 ( DIC  DAV )  3 (OCIC  OCAV )   4 ( PIC  PAV )  
Hypothesis +

+
?
?
Where:
– First term stands for macroeconomic volatility
effect;
– Second for diversification effect;
– Third for microeconomic or banking sectorspecific one; and
– Fourth for bank profitability effect.
21
Exposition:
CARIC  1 ( IC   AV )  .....

Theoretically dCAR=α(0.026-0.0098)= α0.0152
– where: 0.08 is the mandatory capital requirement;
0.26 is the standard deviation of the GDP of Iceland 1989-2000;
0.0098 is the standard deviation of the OECD countries GDP over the
same period;




How to determine or calibrate α?
The value of the parameter α can be treated as adjustment parameter
of the mandatory minimum and it can be derived using cross-country
panel data estimates;
In a logarithmic multiple regression framework, α can be interpreted
as an elasticity;
One can also impose additional restrictions to different parameters
or
22
group of parameters in order to determine their relative importance.
IX. Financial Stability Considerations

Two-front approach:
– Accounting for structural vulnerabilities;
– Improving risk management practices.


Structural vulnerabilities need to be located in the main
sectors of the economy and assessment to be made as to
how these can be met with corresponding buffers;
Concerning risk management practices, the optimal
trade-off needs to be found concerning market induced
vs. regulatory imposed risk management practices, both
at the banking industry level and the individual firm
level.
23
Structural Vulnerabilities and Buffers
Sectors of the
Economy
Real Sector
Sources of
Vulnerability
Profit variability
Own capital
Financial Sector
Market Volatility
CAR
Contingent
Liabilities
Business Cycle
Systemic Risk
Currency
Fluctuations
Sustainable
Budget Surplus
Public Sector
External Sector
Buffers
FX Reserves
24
Market-Based vs. Regulatory-Imposed Risk
Management Practices
Regulation-Based
Market-Based

Predominant role of:
– Comprehensive risk
management practices
within the banks;
VS.
– CAR, VAR, Credit Risk Modelling
etc;
– Banks and MUST;
– Credit Rating Agencies Indications
– Transparency of market information
and active shareholders;
– Mandatory rolling-over of certain part
of sub loans on a secondary market.

Predominant role of:
– FSA power;
– Central Bank Regulations;
– Accounting and Risk
Management Standards;
– Other government regulation;
25
The Early Experience of the Old Basle
Capital Adequacy Framework:
1988
26
1992
27
1996
28
Model of Bank Behaviour
Under Capital Regulation
max   r (L)  L  r
L
B
 B  rD  D  rR ( R)  R  C( BIS)
L, B, R
s.t.L  B  R  D  K
• Variable BIS is defined as:
BIS 
KR
L
L - commercial loans; B - bonds; D - deposits; R - subordinated
debts; K - capital; C(.) - cost function of BIS; r - interest rates;
rR '  0, rR ' '  0, rL '  0, rLL ' '  0
C '  0, C ' '  0
29
Nominator vs. Denominator
Changes of the CAR in G-10
Number of cases where changes in capital and risk-weighted
assets contributed positively (+) or negatively (-) to the
change in capital adequacy ratio
Risk-weighted assets
Capital
Total
+
-
Total
+
18 (19 %)
70 (73 %)
88 (92 %)
-
5 (5 %)
3 (3 %)
8 (8 %)
23 (24 %)
73 (76 %)
96 (100 %)
30
Areas of Impact
of the Basle Capital Framework

Impact on Bank´s Balance Sheets:
– Level of Capital Ratios;
– Structure of Capital;
– Risk-taking Behaviour;


Capital Arbitrage Effects;
Real Sector Effects:
– Impact on Net Domestic Credit;
– Impact on Output;

Impact on Long-Run Competitiveness of Banks:
– Banks vs. Other Credit Institutions;
– International Competitiveness Considerations
31
Contrasting Empirical Evidence






In adjusting their balance sheets, banks attempted to
respond in the least costly way to binding capital
constraints, depending on the cycle and the financial
position;
Large and growing capital arbitrage may be motivated by
other factors, such as taking advantages of economies of
scale, better diversification of funding sources etc;
Changes in bank capital affect lending;
Money matters vs. lending matters for output growth;
Evident shift in the funding share by type of agents that can
not be attributed solely to the capital regulation;
Cost of capital matters.
32
The Real Meaning of the
Capital Adequacy Ratio

The Fundamental Rule:
Primary SecondaryCapital
 8%
Risk - weighted Assets  Off - balance sheet items

Using the Basle model should not mean that the system
is faithfully copied or not, but whether the appropriate
adaptations were made to reflect local conditions;
– Adaptations should be accounted for the risk environment
– Ratio analysis should always be complemented by qualitative
assessment of the bank´s ability to manage its risks;
33
Other Aspects Deserving Particular
Attention:
Reliable market pricing of assets,
particularly loan portfolio review and assets
classification;
 Collateral (re)valuation;
 Loan loss provisioning and all that;
 Interest income recognition policies;
 Volatility and deepness of the markets of
operation;

34
Iceland Commercial and Savings Banks:
CAR and the Risk Profile
Iceland: Capital Adequacy Ratio, GDP Growth Rate
and Lending Growth 1989-2000
%
30
25
20
15
10
5
0
-5 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
Lending Growth
GDP Growth
CAR
35
CAR, Macroeconomy and Banking
Sector Indicators: Regression Results
Does CAR reflects the macroeconomic
development and how macroeconomic
shocks are accounted for?
 The impact of CAR on bank lending: has
Icelandic banking sector experienced credit
crunch as a result of the imposed capital
constraint;
 Structural dimension of the CAR: which
parts tend to add to the rise/fall in the CAR?

36
Variables Included
– GDP - Real GDP growth rage, as proxy for the overall
macroeconomic stance and the demand side of the bank lending;
– BLEND - Annual rate of change in bank lending;
– BDEP - Bank deposits as share of GDP;
– GRSAV – Gross saving rate, annually, as supply side factor of the
bank lending;
– DISC – Discount Rate, as proxy for the supply side of the bank
lending;
– SUBL – Subordinated Loans;
– TTRADE - Terms of trade, as proxy of the macroeconomic
fluctuations;
– DSHOCK - Dummy variable, having a value of 1 in the years
when economy was hit by shock, and zero otherwise, as proxy for
37
the shock performance;
Results

Dependent Variables: CAR (A1); BLEND (A2; A3)

Independent Variables:
A1
A2
A3
GDP t
GDPt 1
BDEPt
BDEPt 1
-0.03
-0.01
-0.31
0.62
0.04
-0.33
(0.17)
(0.07)
(2.34)
(4.48)
(0.32)
(0.38)
DISR
CAR
GRSAV TTRADE DSHOCK
1.25
1.40
-0.07
1.42
-20.87
(1.12)
(2.43)
(0.15)
(1.91)
(1.82)
0.99
0.78
-0.02
0.05
0.75
-8.36
(2.06)
(2.61)
(0.01)
(0.07)
(1.85)
(1.58)
R2
0.55
0.68
0.89
38
Results Interpretations



Higher GDP growth rates and positive terms of trade
have profound demand effects on lending causing
increase in loans where excessively deteriorating
macroeconomic conditions reduce credit;
Sign of discount rate is positive contrary to the theory
where higher the discount rate the lower should be the
rise in credits;
Increase in lending is accompanied by deteriorating
capital adequacy ratios;
39
Cont.



On the other hand, capital ratio moves
anticyclically w.r.t. the movement of GDP growth
rate, while at the same time being negatively
affected by the macroeconomic shocks;
Crucial issue is how banks are achieving higher or
maintaining the desired level of the CAR by not
reducing lending, for example?
CAR level is approached in a residual fashion by
the changes in the other parts of the profit
function.
40
Sub-Loans and Bank Capital
Iceland: Subordinated Loans and
Banking System Capital
1995-97
25
ISK Bn.
20
21,7
20,5
19
15
10
5
2,1
3,5
4,3
Banking System Capital
Excluding Subordinated
Loans
Subordinated loans
0
1995
1996
1997
41
Iceland: CAR and Sub-Loans
Iceland: CAR with and without Subrodinated Loans
1995-98
7,8
1998
8,1
1997
9,8
9,8
9,2
1996
10,7
10
1995
0
5
CAR Excluding
Subordinated Loans
CAR
10
11,1
15
Percent
42
The Impact of the
New Basle Capital Adequacy Framework

Points for Discussion:
– Size versus distribution of the CAR;
– Procyclicality versus anticyclicality;
– Banks behaviour versus CRAs behaviour;
43
Conclusion
CAR is becoming one of the main financial
indicators along with the prevailing market
conditions that always require certain margin
over the minimum;
 If banks should set more capital aside as the risk
increases, how can one prevent sharp portfolio
reallocations in times of stress?
 CRA´s behaviour will be crucial for the financial
stability.

44