DIFFRENTIAL PREMIUM ASSESSMENT SYSTEM

1
DIFFRENTIAL PREMIUM ASSESSMENT SYSTEM
Regional IADI NDIC Conference Deposit
Insurance in Africa Issues, Challenges and
Prospects Venue Central Bank of Nigeria,
Auditorium Central Business District 21 23 June
2004
Presented by A. Sinha
Chief General Manager
DICGC INDIA
2
2
Introduction Differential Premium System
refers to a system of charging premium on the
basis of banks risk profiles as contrasted to a
flat rate premium system. Advantages of
Differential Premium System Relieves moral
hazard i.e. helps curb excessive risk taking and
provides good incentive to banks to manage their
risks. Fairer to the system. Subsidy from
stronger banks to the weaker banks is reduced
(difficult to eliminate subsidy altogether).
3

• Complexity in introducing Differential Risk
  Assessment System
• Risk measurement and pricing of premium is a
  complex task.
• Difficulties in finding appropriate and
  acceptable methods of
• differentiating banks risk.
• Obtaining regular and timely data.
• Resource Constraints - Availability of skills
  and IT systems.
• Ensuring transparency of rating criteria.
• Containing the potential destabilising effects
  of imposing
• high premia on troubled banks.

3
4
Trends in Transition to Differential Premium
Assessment System

• Logically attractive - increasingly larger
  number of countries are moving / contemplating
  to move to Differential Premium Assessment
  System.
• CDICs International Deposit Insurance Survey
  2003 15 countries currently have in place
  Differential Premium System.
• IMF Occasional Paper 197 (2000) Deposit
  Insurance Actual and Good Practices by Gillian
  G.H. Garcia placed this number at 24 i.e. more
  than 1/3rd of the countries with explicit deposit
  insurance scheme.

4
5

• International Guidance Documents
• IMF Occasional Paper 197 (2000) titled Deposit
  Insurance Actual and Good Practices by Gillian
  G.H. Garcia
• Financial Stability Forums Guidance for
  Developing Effective Deposit Insurance System
  (September 2001).
• IADIs Draft of April 2004 on General Guidance
  for Developing Differential Premium System.

5
6
Key Points of Guidance in IADIs draft on General
Guidance for Developing Differential Premium
System Objectives of Differential Premium System
To provide incentive for banks to avoid
excessive risk taking and introduce more fairness
into the premium assessment system. ? Deterrence
effect of insurance premium may be insufficient
for weak banks - Needs to be supplemented by
early warning system and prompt corrective action
(PCA) as a part of a sound regulatory and
supervisory regime which together with Lender of
Last Resort facility and an incentive compatible
Deposit Insurance System forms an efficient
financial Safety Net for banks.
6
7
? Provision to terminate deposit insurance cover
in extreme cases. As per FDICs survey of 73
foreign deposit insurance organisations in
January 2000 ( Risk Assessment Results on an
International Survey of Deposit Insurers FDIC
Banking Review 2003, Vol. 15 No.1 ), of the 36
insurers who responded to this issue, 11
confirmed having this power but only 2 had ever
used this authority.
7
8

• Situational Analysis against Conditions
• As a Differential Premium System is a complex
  system, it is important to undertake a
  situational analysis to assess the state of
  economy, current Monetary and Fiscal Policies,
  the state and structure of the banking system,
  public attitudes and expectations, the legal
  framework, the strength of prudential regulation
  and supervision and the soundness of accounting
  and disclosure regimes.
• Gaps between existing conditions and more
  desirable conditions should be identified and
  available options should be thoroughly evaluated.

8
9

• Approaches used to differentiate bank risk
• The approach should (1) be effective at
  differentiating banks into appropriate risk
  categories (2) utilise a wide variety of relevant
  information (3) be forward looking and (4) be
  well accepted by the banking industry and
  financial safety net participants.
• Sources of information/data
• Quantitative Audited balance sheets,
  off-site returns, examination findings and market
  data for large banks (equity prices, interest
  rate or yields on inter bank deposits,
  subordinated debt and debenture)
• ? Qualitative
• Regulatory and supervisory CAMELS or similar
  ratings which also incorporate judgement on a
  number of other criteria like key business
  practices, banks corporate governance, strategic
  management, internal control, risk management and
  compliance etc.

9
10

• Approaches to risk profiling of banks for
  premium assessment purposes.
• Quantitative criteria approach
• Use of one or a combination of quantitative
  factors - capital adequacy (most common),
  quality and diversification of asset portfolio,
  sufficiency, volatility and quality of earnings,
  stability and diversification of funding and risk
  exposures, etc.
• Risk differentiation methodology
• Expected Loss Pricing (i.e. KMV, Multivariate
  Discriminant
• Analysis Probit and Logit Models, etc.)
• Option theoretic approach (Merton, Ronn Verma)
• Advantage Objective

10
11

• Drawbacks heavily dependent on high quality,
  consistent, reliable and timely data, complex
  methodologies, availability of market data only
  for large banks in countries with well developed
  capital markets.
• Accounting data are lagged data.
• ? Qualitative Criteria
• Advantage Provide some information on
  current and future risk profile of banks and
  hence forward looking.
• Disadvantage Less transparent, highly
  judgemental. May result in large number of
  appeals by banks and more difficult to defend.

11
12
? Combined quantitative and qualitative criteria
approaches Most common Differential Premium
System. Several countries - Argentina, Canada,
France, Taiwan and the USA utilise this approach.
Determination of inter-se weightage to the
quantitative and qualitative factors is a major
issue. The two best known such approaches
are that of FDIC (USA) and CDIC (Canada). In
USA, banks are slotted into a 3 x 3 matrix with
an equal weightage to the two criteria adopted
i.e. capital adequacy and supervisory ratings.
CDIC uses a score card incorporating 14
individual quantitative and qualitative measures
with 60 per cent weightage to quantitative
indicators. Based on scores, banks are slotted
into 4 premium categories.
12
13

• In Argentina, a basic premium is charged to all
  banks with additional premium levelled by a
  combined qualitative ( CAMEL rating
  ) and quantitative (excess or deficiency of
  capital over the required minimum and the quality
  of the loan portfolio) criteria.
• Achieves risk differentiation, not risk
  determination.
• Advantage - uses widest range of information to
  assess banks risk profile.
• Drawback - may impose a higher information
  requirement on banks and could be more open to
  challenges compared to approaches using mostly
  quantitative criteria.

13
14

• Authority, Resources and Information Requirements
• Deposit Insurance Authority (DIA) should have
  the necessary authority, resources and
  information in place Where DIA relies on
  supervisors for information, a formal agreement
  needs to be in place.
• Information requirement should not be unduly
  burdensome to banks.
• Information needs to be validated by
  establishing reporting standards and on-site
  verification. Previously audited information
  would be useful.
• Most recent data should be used. Since frequent
  updating is not practical, usually fiscal year
  end audited financial information is used for the
  risk profile determination for a year.

14
15

• If the DIA provides insurance to more than one
  category of financial institutions, it will be
  desirable to devise assessment methodologies to
  suit the risk profiles of different categories of
  member institutions.
• Even within one category i.e. banks, there may
  be a need to devise different assessment
  methodologies particularly for (a) new banks and
  (b) large banks.
• ? New Banks Risk profile would be different as
  these operate with unproven business plan, have
  high capital asset ratios and low level problem
  assets compared to established banks.
• ? Large Banks Special characteristics much
  lower level of core funding, complex operations,
  dynamic risk profile, availability of valuable
  market data and information, etc.

15
16

• Premium Categories and Assignment of Premium
  Rates
• Number of premium categories differs across
  various deposit insuring agencies - USA and
  Taiwan have 9, CDIC has 4. Argentina and France
  use a continuous premium function linked to the
  risk profile of banks.
• Number of premium categories to be decided
  keeping in view the size and number of banks to
  ensure meaningful distinction between premium
  categories to act as an incentive for banks to
  improve their risk profile.
• Large number of premium categories would help
  in risk differentiation, but would reduce the
  incentive for banks to move to the lower premium
  category.
• Too small a number would result in clubbing
  banks with significantly different risk profiles
  into one group.

16
17

• Premium rates should be set to ensure that the
  funding requirements of the deposit insurance
  system are met over the given time horizon and to
  provide effective incentive to sound risk
  management by banks. It should be ensured that
  even the best rated banks pay some premium as
  they also pose some risk to the deposit insurance
  fund.
• Underpricing would increase moral hazard and
  affect the solvency of the DIA.
• Overpricing would reduce resources of the banking
  system.

17
18

• Transition
• A transition plan should set out the transition
  objective, responsibilities, resource
  requirement, time table and deliverables. The
  plan should be communicated to all interested
  parties well in time.
• Greater adjustment period would be required for
  more complex Differential Premium Systems.
• Approach to deal with the potential
  destabilising effects of higher premiums on
  already troubled banks needs to be devised. For
  example, the Differential Payment System could
  be implemented in stages with advance warning of
  when and how of the different stages. A
  transition period where virtually all banks
  receive favourable treatment to place themselves
  in low premium categories could also be
  considered. The advantage would be of reducing
  the initial impact of premium increase
  particularly for troubled banks but provide
  incentive to them to improve their category
  rating over time.

18
19

• What happens if there are banks after the
  transition period for whom the premium would be
  destabilising ?
• Note An interesting transition mechanism - have
  flat rate premium, but link
• to Risk weighted Assets, not deposits.

19
20

• Transparency Disclosure and Confidentiality
• Criteria used in differential premium system
  should be transparent to banks and other
  participants.
• An appropriate balance should be struck between
  promoting accountability, discipline and sound
  management through disclosure and the need to
  ensure confidentiality of information. Disclosure
  of premium categories can enhance market
  discipline but can have negative consequences
  also.
• Extent of disclosure is an evolving policy
  issue - over a period of time disclosures have
  increased substantially. While there is little
  disclosure of supervisory information there is a
  debate on disclosure of supervisory information
  to the public.

20
21

• Currently some systems have adopted a policy of
  partial transparency for example, USA, Canada,
  Taiwan disclose basic framework of the system
  and the factor criteria to the public but the
  actual ratings or premium categories are
  disclosed only to the Board of Directors and
  Management of the banks. Banks are prohibited
  from disclosing this information.
• At present, no deposit insurance system
  publishes these ratings.

21
22

• Review, updating and fine tuning of differential
  premium system
• A formal process to review potential
  disagreements regarding risk categorisation for
  premium purposes should be implemented to resolve
  disputes.
• Differential Premium Systems need to be
  regularly reassessed on their effectiveness and
  efficiency in meeting their objectives and if
  necessary, updated and / or revised to changing
  conditions or requirements i.e. implementation
  of Basel II.

22
23

• Premium system in USA
• In the light of large scale bank failures in
  1980s, risk based premium system introduced with
  effect from January 1, 1993.
• Banks are rated into a 3 x 3 matrix based on 3
  classifications for capital ratios (well
  capitalised, adequately capitalised and under
  capitalised) and 3 categories of supervisory
  ratings ( A, B and C).
• CAMELS ratings of all banks are available every
  12 months ( 18 months for high rated small banks)
  and capital ratios are based on financial reports
  filed quarterly by the banks.

23
24

• Evolution of Premium Rates

First Columns under A, B and C Schedule
Effective January 1, 1993 Second Columns under A,
B and C Schedule adopted in 1995 Third Columns
under A, B and C - Current Schedule
24
25

• One of the objectives when risk based premium
  rate schedule came into effect on January 1, 1993
  was to produce sufficient revenue within 15 years
  to recapitalise the deposit insurance funds that
  had been depleted by the large failure costs of
  the 1980s.
• The lowest premium of 23 cents charged to the
  safest banks was the rate charged to all banks
  prior to introduction of risk based premium
  system.
• High premium to the safest banks and maximum
  differential 8 cents
• - motivated by the need to build up funds quickly
  and to protect the lowest rated banks during the
  transition period.
• Schedule revised when the DRR of 1.25 per
  cent of insured deposits was reached in 1995.

25
26

• Premium rate drastically reduced by 19 cents
  for the best banks. No change for the worst rated
  banks - differential increased from 8 cents to
  27 cents. Premium of 31 cents still much lower
  than the actuarially fair premium.
• Subsequent legislation prohibited charging any
  premium to the safest banks when the DRR gt 1.25
  per cent.
• Now safest banks do not pay any premium.
• Differential between safest and weakest banks
  retained at 27 cents but the premium level for
  weakest banks reduced from 31 cents to 27 cents.
• Consequence of the current schedule - 92 of
  all institutions pay no premium. Currently over
  900 newly chartered institutions have never paid
  any premium.
• Reform of deposit insurance system in USA is
  under way to address, among other things, the
  pro-cyclicality of premium regime, adoption of
  flexible range bound DRR, and to ensure that
  every bank pays some premium.

26
27

• Premium System in Canada
• Risk based premium based on a scoring system
  introduced in 1999.

27
28
28
29

• 60 per cent quantitative and 40 per cent
  qualitative scores.
• Two most important parameters Examiners
  rating - 25 per cent and capital adequacy - 20
  per cent.
• Capital adequacy supplemented by several other
  quantitative criteria (40 per cent) to assess the
  ability of a member institution to sustain its
  capital.
• Premium Categories

29
30

• Differential between category 1 and 4 has
  substantially reduced from 29 basis points to 14
  basis points as a result of premium halved across
  the board.
• The premium at 16 basis points is likely
  significantly less than actuarially fair premium.
• Nevertheless the premium structure meets the
  objective of CDIC.
• To help the weakest banks in particular a
  transitional scoring mechanism operated in the
  first 2 years.
• ? First year - quantitative score of each member
  increased by 20 per cent.
• ? Second year Quantitative score increased by
  10 per cent.
• New banks not having sufficient operating
  history for volatility measures given a score
  based on the average of other quantitative scores.

30
31

• A General Framework for Differential
• Premium Assessment
• Banks are exposed to credit risk from the
  portfolio of loans. Similarly Deposit Insuring
  Agencies (DIAs) are exposed to credit risk from
  the portfolio of insured banks due to the
  possibility of failure of insured banks.
• Distribution of credit losses in a loan
  portfolio is typically skewed towards large
  losses (Exhibit 1). DIA also faces skewed
  distribution of losses High probability of
  small losses from the failure of a number of
  small banks and small losses from the failure of
  a large number of small banks (Exhibit 1).
• Banks provide for expected losses through loan
  loss reserves and unexpected losses through
  Economic Capital. Economic Capital is determined
  so that the estimated probability of unexpected
  credit losses exceeding the economic capital is
  less than some target insolvency rate. In
  practice target insolvency rate is chosen to be
  consistent with desired credit rating.

31
32
In a DIAs portfolio of insured banks ELi PDi
X ExPi X LGDi Most important
component ?i ExPi v (PDi X ?2LGDi LGDi2
X PDi (1- PDi ) EL (P) ? PDi X ExPi x
LGDi  ?(P) v ?? Pij ULi ULj
32
33

• Choose UL(P) at a given insolvency rate or
  desired credit rating.
• Premium to be charged- Expected loss and banks
  contribution to meet the solvency standard
• Expected loss pricing makes fund self financing
  through time and relieves moral hazard. Hence
  premium at the minimum should be equal to the
  expected loss.

33
34

• Expected Loss Pricing Methodologies
• 1.      Balance sheet based indicators
• A sample of banks is selected and classified
  into two groups - failed or not failed.
• On the basis of statistical methods the best
  subset of variables (financial ratios) is
  selected which best distinguishes between the two
  groups within a certain prediction horizon.
• Statistical techniques used - Multiple
  discriminant analysis, logit / probit models
  etc.
• Once the variables and parameters are
  determined, these can be used to predict the
  probability of failure of banks.
• For every bank in the sample
•   ? ?o ? ?x , ? 1 ? gt 0 failed
• 0 ? 0
  not failed

  
34
35

• Xs are explanatory variables like Profitability,
  Leverage, loan growth and ROA, etc.
• is a latent variable which has either standard
  logistic distribution or standard normal
  distribution.
• It can be then be shown that P ? 1 ? ? gt0
  ?(?o S x)
• 0 ? (?o S ? x) 1
• Where ? is a Cumulative distribution function for
  (a) standard logistic random variable (Logit
  model) or (b) standard normal random variable
  (Probit model).
• The Parameters of the function ? (.) are
  estimated from the sample of failed and
  non-failed banks.

35
36

• 2. Market based indicator - Equity Prices -
  KMV's Distance to Default
• Based on Merton's model in which a firm has
  only one liability - a zero coupon bond whose
  maturity value is B.
• The Asset value (v) evolves through a geometric
  Brownian motion.
• Equity holders have a call option on the
  assets of the firm (V).
• If V gt B at maturity, equity holders pay off the
  Bond holders and retain the balance (E V-B) as
  equity.
• If V lt B, the firm is insolvent (E O) and the
  equity holders turn the assets to the Bond
  holders who receive V (ltB).

36
37
? Using Black Scholes option pricing formula
E VN (d1) - BerT N (d2)
(1) and EsE N (d1).Vsv
(2) Where d1 log V
? s2v T B
2
svvT and d2 d1 - svvT T
Time to Maturity N (.) Standard normal
distribution ? risk free rate of interest
sv Standard deviation of V sE
Standard deviation for E
37
38

• E, sE can be observed from market value.
• V, sv are unobservable but can be calculated
  using the values of E and sE in equation (1) and
  (2).
• KMV take trigger for default not when V lt B as
  in Mertons model but when V lt short term Debt
  (STD) 1/2 Long Term Debt (LTD).
• KMV define Distance from Default
• DD V - ( STD 1/2 LTD )
• Vsv
• DD is then mapped into PD using KMV's data base.

38
39
3. Ideally PD determination should be
done using all available data and information -
market, accounting and supervisory. FDIC in its
Final Report dated 16th July 2003 on
"Strengthening Financial Risk Management at the
FDIC" has stated that a statistical model of the
probability of failure for individual
institutions should be developed using CAMELS
ratings, financial metrics from Call Report data
and relevant market data.
39
40

• Option Theoretic Models Direct Computation of
  Premium
• Based on Mertons model and additional insight
  that the insurer effectively writes a put option
  on the Assets of the bank, Premium can be
  directly computed.
• Merton interprets the length of time until
  maturity as the length of time until the next
  audit of the banks assets.
• Ronn Verma (RV) replace the Zero coupon bond
  in Mertons model by insured deposits (face value
  B1) and all other debt (face value B2).
• Total debt B B1
  B2
• Using Block Scholes formula for valuing put
  option, premium per dollar of deposits
• g N (-d2) - V N
  (-d1) ....... (1)
• 
  B 

40
41

• Where d1 log V ?v2 T
  
• B
  2
• ?v vT
•   and d2 d1 - ?v vT
•      Equity is a call option on the assets of the
  bank with strike price B
• E VN (d1) BN (d2) .......
  (2)
• and E?E V?V.N (d1) .......
  (3)
•        E and ?E are observable from the market
  data.
• V and ?V are non-observable.

41
42

• Equations (2) and (3) can be used to value V
  and ?V and use it in equation (1) to compute g.
• RV further built in regulatory forbearance by
  assuming that after the annual audit, the
  regulators will not liquidate the bank unless V lt
  l B where 0lt llt1. l measures the regulatory
  forbearance.
• This will modify equation (2) to
• E VN (d1) lBN (d2)

42