CASE STUDY DEPOSIT GUARANTEE FUNDS

Transcription

1 CASE STUDY DEPOSIT GUARANTEE FUNDS 18 DECEMBER FINANCIAL SERVICES

2 Section 1 Introduction to Oliver Wyman

3 Oliver Wyman has been one of the fastest growing consulting firms over the last 20 years Key statistics Industries 2012 Revenue: $1.5 BN CAGR: 10% $2,0 BN $1,5 BN Revenue $1,0 BN $0,5 BN $0,0 BN

4 Examples of recent project experiences in the Nordic region Credit process redesign Performance and strategy review Oliver Wyman supported a major Nordic bank in redesigning the credit process for small and medium-sized corporates The objectives of the project was to Increase efficiency in the credit process through standardisation and automation Reduce headcount Increase harmonisation in credit decision making The project designed a new end-to-end credit process, including Credit decisioning Risk scoring Pricing A European bank that experienced a significant drop in profitability and ROE during the financial crisis asked for Oliver Wyman s support in improving its financial performance Oliver Wyman performed an assessment of the situation and developed a range of recommendations that the client should take to improve its performance All initiatives were prioritised and detailed in an implementation roadmap, taking into account Regulatory impact Business value Cost of implementation Organizational constraints (e.g. resource capacity, infrastructure, capabilities, etc.) 3

5 Section 2 Case study Deposit guarantee fund

6 Deposit guarantee systems protect bank depositors from losses during bank failures and thereby work as a safety net and promote financial stability What it is Deposit guarantee schemes reimburse deposits to depositors whose bank has failed From the depositors point of view, this protects a part of their wealth from bank failures From a financial stability perspective, this prevents depositors from making panic withdrawals from their bank, thereby preventing severe economic consequences Why it is needed Gives depositors comfort that their funds are not at risk, thus reducing the risk for bank runs Reduces volatility among depositors as they do not need to assess bank riskiness when depositing money Provides countries with an orderly process for dealing with bank failures and a mechanism for banks to fund the cost of failures How it works Deposit guarantee systems are most often established and managed by a government body Should adhere to a number of principles Membership should be compulsory for all financial institutions that accept deposits Coverage in the EU is currently EUR Pay-out time in case of bank failure is typically 7 14 days Source: European Commission, BCBS 5

7 A finance minister asked Oliver Wyman to support the structuring of the country s deposit guarantee scheme The project was delivered in three stages 1 Hypothesis generation Diagnostic 2 3 and analysis Provide recommendations Understand the country s specific rules and regulations Understand the client situation and context Understand how other countries have designed their deposit guarantee frameworks Based on the outcome from the diagnostic, develop a hypothesis for the target structure of the guarantee scheme Discuss hypothesis with client to get input and challenge Conduct analysis to test the hypothesis Provide recommendations for a target structure based on the findings from the analysis Develop an implementation roadmap for the client 6

8 Deposit guarantee frameworks are typically set up in one of two ways Description Pros Cons Public treasury 1 2 Taxpayer money is used to pay out to depositors in case of bank failure Simple system as the government pays out from public treasury in case of emergency Not a popular approach with the public as taxpayer money is used to reimburse depositors Banks don t bear the costs for potential failures Reactive approach that draws funds from public treasury in case of emergency rather than proactively building up funds Guarantee fund A fund is set up which can be used to draw funds from in case of bank failure All banks that hold deposits must pay in to the fund Banks bear the risk of failure themselves Popular approach with the public as it doesn t burden public treasury Proactive approach as it collects funds before they are needed More complex system than pulling funds from the public treasury; a number of practicalities must be agreed, e.g.: How much should banks pay in to the fund each year? What should the target size of the fund be? Should the fund take risk with its assets? Selected project approach 7

9 Banks use an Expected Loss framework to estimate losses due to e.g. loan impairments Expected Loss (currency) Probability of default (%) Loss given default (%) Exposure at default (currency) The estimated loss on a portfolio The average percentage of borrowers that default in the course of one year The magnitude of likely loss (the percentage of the total loan amount that the bank would lose) in case of default The outstanding amount to which the bank is exposed to at the time of default Example: If Nordea lends out 100 SEKm to Ericsson, the expected loss on that loan is SEK if: PD = 1% (there is a 1% risk that Ericsson defaults within one year) LGD = 50% (if Ericsson defaults within one year, the loss for Nordea will be 50% of the loan size) EAD = 8 SEKm (the loan amount at the time of default is estimated to be 8 SEKm) 8

10 Deposit guarantee funds can use the expected loss framework to estimate a theoretical size of the fund Banks Expected Loss For a guarantee fund, the expected loss is the amount that is expected to be drawn from the fund Probability of default For a guarantee fund, PD is the probability of default for a bank in the financial system Loss given default For a guarantee fund, LGD represents the proportion of the exposure that must be drawn from the fund (In case of default, any assets available in the bank are used first, before the fund is tapped) Exposure at default For a guarantee fund, EAD represents the total amount of insured deposits at the time of default What are the weaknesses with using an Expected Loss framework to estimate the target size of the fund? 9

11 Example of PD estimation: Both qualitative and quantitative measures are evaluated to calculate the probability of default Example process for calculating the probability of default (generic example) Probability of default Qualitative factors Quantitative factors Financial strength of owner Experience of leadership Number of missed payments Total sales ROE Debt-to-equity ratio Regress using historical data A regression model is built to find statistical relationships between PD and the chosen quantitative and qualitative factors Expert judgment overlay Ultimately, expert judgment is applied to any model output to cater for risk aspects not captured by the model 10

12 Probability of default is inherently difficult to estimate in practice as it is a measure of future events Some of the problems with estimating PD are: Most estimates are based on historical experiences, which may or may not be representative of the future There are correlation effects that are difficult to assess Models may be overly simplistic and make crude assumptions of reality Available information regarding the borrower or institution is not always available, correct or sufficiently detailed Etc. Due to these reasons, many institutions complement their PD models with stress testing, e.g.: Scenario analysis Sensitivity test of input parameters 11

13 There are different types of stress testing approaches Scenario tests Sensitivity tests Historical scenarios Actual historical events are applied to the portfolio to assess implications The purpose of the test is to assess how the portfolio would perform during these scenarios 9/11 terrorist attacks Asian Crisis 1997 Black Monday Oct Hypothetical scenarios Hypothetical scenarios are created to stress the portfolio, e.g.: Banking crisis leading to 10% GDP decrease and 10% increase in income taxes Housing crisis leading to 10% year-on-year property price decreases and doubling of unemployment over a five year period Hypothetical scenarios can be tailored to be more relevant to the portfolio and current market environment than historical scenarios In sensitivity tests, underlying variables are adjusted to investigate portfolio impact, e.g.: Doubling of PD in portfolio for certain customer segments 20% reduction of deposit volumes Sensitivity tests are typically easy to define and implement and are often used at trading desks Reverse tests try to identify how large shocks are required to trigger a predefined event, e.g.: How much must the stock market fall for us to make a 10% loss on the portfolio? 12

14 How much should each bank pay into the fund? The project proposed a fee structure based on three key drivers Size of insured deposit base 1 Forecasting of future deposit base Proportion of insured vs. uninsured deposits 2 3 Bank riskiness (probability of default) E.g. using banks credit ratings (from S&P or other rating agency) Base bank riskiness on amount of capital held for risky assets (according to Basel regulation) Allowed time for fund to reach target size The longer time it takes for the fund to reach its target size, the lower the banks fees will be Bank specific fee Fair system for participating banks as the fee takes into account bank specific characteristics The system results in cheaper deposit funding for safer banks 13

15 Should the fund take risks with its assets? Should the fund invest its assets and aim to get a return? Pros Cons The target size of the fund is (hopefully!) reached faster Participating banks can reduce the size of their fees Risk that size of assets is reduced if investment decisions are poor There is a political risk associated with this strategy and poor investment decisions may result in a political crisis The construction of the guarantee fund led to asymmetric risks that resulted in banks wanting to take more risks with the assets in the fund: If the fund would perform well, banks would reduce their fees If the fund would perform poorly, banks wouldn t have to pay higher fees If banks would default, they wouldn t have to care about the size of the fund; the state would cover any shortcomings of the fund 14

16 On-going management of the fund: Like any other business, guarantee funds are exposed to a wide range of risks Major risk types for the fund Liquidity Risk Market Risk Concentration risk Interest Rate Risk Credit Risk Operational (Event) Risk Description/examples Will there be a market to sell the assets in? Will the assets lose value due to unfavourable market movements? Will specific losses in one asset have a large impact on the total portfolio? Will the assets lose value due to unfavourable interest rate movements? Will a counterparty default on its obligations? Will we encounter losses due to unforeseen events regarding people, process or system failures? 15

17 If the fund decides to invest its assets, a number of complications need to be addressed Complication How much risk should the fund take with its assets? Implication The fund should agree the level of accepted risk and tolerated losses with relevant stakeholders How can the fund ensure reimbursement to depositors in seven days in case of bank failure? Should the fund support its country s economy by investing in local government bonds? If the fund invests its assets, which risks is it exposed to and how should those risks be managed? The fund should invest in only highly liquid assets that can be sold quickly in case of bank failure If it does, the correlation between country risk and the banking system should be duly assessed and mitigated Market risk (overall performance of financial markets) Liquidity risk (difficulty to sell assets quickly) Credit risk (risk that a counterparty defaults) Etc. All of these issues should be addressed in a Risk Appetite Framework that defines the level of risk that the fund accepts and how they should be mitigated 16

18 Risk Appetite serves as one of the most important mediums for coordinating risk-taking activities across an organisation Strategy and stakeholder expectations Owners return expectations Growth objectives Regulatory risk management expectations Board defined top-down Risk Appetite statement Strategic Liquidity Market Reputational Cascading appetite / linkages to limits Other Credit risk types Embedding Planning and budgeting Capital allocation Incentives and compensation Governance Reporting and tracking New business limits Liquidity risk limits Bottom-up limits framework Market risk limits Reputational risk limits Other Credit Credit Risk limits Early warning indicators Breach management 17

19 Risk Appetite statements should be clearly defined and link to potential management actions that can be used to adjust the risk profile Client example Disguised client example of range of possible Risk Tolerance statements Metric Illustrative definition Green Amber Red Concentration risk We should not have any single exposure >10% of total assets 7% 7% - 10% 10% Quantitative Qualitative Debt rating Our senior debt should at all times stay above a Moody s rating of Baa Earnings volatility We will not miss consensus earnings forecast by more than X % at a YY % confidence level We will aim to consistently target dividend of XXX Maximum loss We do not wish to see a loss of more than XXX at the YY % confidence interval Liquidity headroom Available liquidity resources to meet requirements at XX % confidence interval Reputation Ensure that the highest ethical standards are followed at all times Regulation Have no significant instances of regulatory breach Governance Ensure appropriate policies and processes are followed at all times Growth All new business opportunities to follow appropriate risk controls 18

20 A monthly process should be in place to monitor potential breaches of risk appetite limits Illustrative example of how triggers can be used Projection Capital adequacy ratio Projection indicates imminent breach, requiring the development of an action plan to avoid a breach Warning limit Risk appetite limit Time If ratio enters red territory, the Board must take immediate action 19

21 Section 3 Appendix Rating model build

22 Appendix: How rating models work Rating models contain an objective and expert based part Objective calculation Expert judgement application Capital structure Financial factors Liquidity Profitability Financial score Behavioural factors Limit utilisation Credit turnover Behavioural score Model calculation and final model score Expert override Committee review Final rating Nonfinancial factors No of directors Age of business Non-financial score 21

23 Appendix: How rating models are built The model build typically follows five general stages Multi-factor Generation of Single factor analysis and Model factor long-list analysis model testing selection Calibration A long list of potential model factors is generated Each factor should be believed to be a good indicator of credit risk The long list should typically includes Financial factors Non-financial factors Behavioural factors All factors on the factor long list are examined individually and each factor s correlation to credit risk is assessed Factors that show strong statistical predictability are taken to the next step; poor performing factors are removed Factors are combined and their correlation with credit risk is examined as a combined set of factors An algorithm tests all combinations of factors and factors are given weights using various statistical methods A favourite model (combination of factors) is chosen to be taken to the next step The selected model is tested using various techniques: Out-of-sample testing (test on sample that was not used in model build) Segmentation testing (test model performance on specific customer types, industries, etc.) Plus other types of testing Previous steps result in rank-ordering of borrowers In this step, borrowers are segmented into risk classes and each risk class is assigned a PD 22