Risk Management

  1. Introduction
  2. Commercial banks play an important role in the functioning of any economy and are a significant catalyst for growth. However, banks are highly leveraged organizations. They accept deposits from the public and take risks while creating financial assets. They are exposed to various kinds of risk like credit risk, market risk, operational risk and liquidity risk. The purpose of this chapter is to explain in detail the various risks faced by a bank, outline methods for risk management and examine the regulatory framework. Treasury managers need to be aware of the various kinds of risks, should be able to design risk mitigation measures, devise hedging strategies and comply with the regulatory framework.

    The chapter also highlights some risk management tools that banks can use for risk management. These tools are also subject to regulatory guidelines and a reading of the Basel guidelines/framework or the Reserve Bank of India guidelines is suggested.

    The aim of the chapter is to enable the reader to

    1. understand the various risks that a bank faces like credit risk, market risk, operational risk and liquidity risk
    2. understand the relation between risk and capital
    3. estimate the probability of default
    4. to find out the level of economic capital for a bank
    5. understand the implications of prudential norms like income recognition, asset classification, provisioning and capital adequacy
    6. understand risk management techniques followed by banks
    7. understand the various reasons behind assets turning non-performing
    8. develop appreciation of the Basel guidelines.
  3. What is risk?
  4. Before getting into risk management in banks, the chapter will start with definition of risk and what it means. The difference between risk and uncertainty will be discussed. Risk management is only possible if risk can be identified, quantified and measured. Then only it can be controlled. There cannot be any activity without risk. Question is, how best to control it. It is also important to understand the relationship of credit rating with risk.

    Risk arises when there are various outcomes of an event, and where there is a probability associated with each outcome. In tossing a coin or throwing a dice, the outcomes are known with certainty, like head or tail and 1 to 6. However, which side of the coin will show up after the coin lands, or which number will show up after a dice is rolled, has a definite probability associated with it.

    Uncertainty, on the other hand, is a situation where the outcomes are known, but it is difficult to assign probabilities. That is why there is insurance for protection. There cannot be any insurance for risk. Activities like horse racing, stock market investing, gambling etc. cannot be insured. However, insurance is available for mudslides, air travel, earthquake, death, illness etc. In case of uncertain events, neither can definitive probability be assigned to the outcomes, nor can the extent of impact be forecast. That is why, expenses for medical treatment can be insured, but with limits.

  5. Various Kinds of Risks a Bank is Exposed to - Credit Risk, Market Risk, Operational Risk & Liquidity Risk
  6. Credit Risk

    Credit risk arises when a borrower fails to repay either interest, or principal, or both. Banks accept deposits from individuals and on lend these funds for investment purposes. Borrowers of these funds are at times unable to pay interest, or repay these loans on time. This default risk is known as credit risk.

    Market Risk

    Banks also deploy deposits received from customers in traded securities and financial assets. Examples would be bonds, government securities, treasury bills and currency. The risk banks face due to any deterioration in value of these assets due to changes in market price is called market risk. For example if the rate of interest rises, then there is a fall in the value of the debt securities held. If the there is an adverse movement in the exchange rate, then the value of currency holdings fall. These are examples of market risk.

    Operational Risk

    Any financial loss to a bank arising from lapses in systems and procedures is called operational risk. It can arise from violation of internal codes of conduct leading to fraud or overextension of exposure. It can arise from customer data deletion due to fire, or any natural calamity or breakdown of the IT infrastructure.

    Liquidity Risk

    Against deposits of customers, banks hold financial assets on their books which generate returns. At times, banks are unable to sell such assets to meet liquidity requirements of depositors, as there may not be any buyers. Further, there are situations when banks knowingly do not sell as they feel that any effort on their part to sell will lead to reduction in prices of these assets and they will not get any value. This is liquidity risk. It is a risk arising out of the inability of the banks to sell assets in the market to generate liquidity.

  7. Relation between Risk and Capital
  8. Consider the following tables.

    Table 1

    Funding pattern

    Rs.

    Lending rate 6% pa

    Probability of default

    Equity

    10

    0%

    5%

    10%

    15%

    Debt (borrowing rate 5% pa)

    90

    Return

    106

    95 + 95*.06

    = 95+5.7 = 100.7

    90+ 90*.06 = 90+5.4 = 95.4

    85 + 85*.06 =

    85 + 5.1 = 90.1

    Total

    100

    Repayment of debt with interest

    90*1.05 = 94.5

    94.5

    94.5

    94.5

    Surplus

    106 – 94.5 = 11.5

    100.7 – 94.5 = 6.2

    95.4 - 94.5 = .9

    90.1 – 94.5 =

    -4.4

    Table 2

    Funding pattern

    Rs.

    Lending rate 6% pa

    Probability of default

    Equity

    5

    0%

    5%

    10%

    15%

    Debt (borrowing rate 5% pa)

    95

    Return

    106

    95 + 95*.06

    = 95+5.7 = 100.7

    90+ 90*.06 = 90+5.4 = 95.4

    85 + 85*.06 =

    85 + 5.1 = 90.1

    Total

    100

    Repayment of debt with interest

    95*1.05 = 99.75

    99.75

    99.75

    99.75

    Surplus

    106 – 99.75 = 6.25

    100.7 – 99.75 = .95

    95.4 – 99.75 =

    - 4.35

    90.1 – 99.75 =

    -9.65

    In both the tables, total amount of Rs.100 is being deployed in the market at a rate of interest of 6% pa. In Table 1, Rs.100 is being raised by of equity capital of Rs.10 and debt (deposits in the case of a bank) of Rs.90 at a rate of interest of 5% pa. In Table 2, Rs.100 is being raised by of equity capital of Rs.5 and debt of Rs.95 at a rate of interest of 5% pa. If debt cannot be repaid with interest, then the bank itself will be a defaulter in the market. If the bank is unable to repay the depositors even their deposit amount, then the bank will have to declare bankruptcy.

    Observe that

    1. In table 1, debt can be repaid with interest up to a probability of default of 10%.
    2. In table 1, only principal can be repaid with a probability of default of 15%.
    3. In table 2, debt can be repaid with interest up to probability of default of 5%.
    4. In table 2, only principal can be repaid with interest with a probability of default of 10%.
    5. A higher level of equity funding can absorb a higher level of default.

    The probability of default of a borrower increases with the riskiness of the venture of the borrower. However, rates of return of risker ventures are also higher than those of less risky ventures. Thus, to increase returns, banks have to acquire some risky assets in their portfolio. The above two tables show that, in order to earn higher returns from riskier ventures that have higher probability of default, banks need to have higher equity capital to absorb the incremental risk. Thus, the moral is: have capital, take risk.

    (The above tables assume the same rate of return from less risky and more risky ventures. The reader can perform their own exercise with higher rates of return with higher default rates.)

  9. Expected Loss
  10. Expected Loss (EL) is defined as

    EL = Probability of Default (PD) * Exposure at Default (EAD) * Loss Given Default (LGD)

    EAD is the amount of loan outstanding of the borrower at the time of default. Generally, loans are advanced against collateral or security. For example, term loans are secured by mortgage on land and building and hypothecation of plant and machinery. Working capital loans are secured by hypothecation of current assets. At the time of default, the assets of the company will have some value. The total outstanding loan, minus the recovery value, in % is LGD. PD is the probability that a loan account will default. This is related to the credit rating of the loan. It can be calculated as the ratio of total loans defaulting in this category to total loans in this category. Not every loan in a BBB credit rating category defaults. Again, not all loans in the AA category pay. Probability of default of a loan is the ratio as defined above.

    Consider a numerical example. Let Rs.100 be the amount of loan outstanding. Let 60% be LGD and 3% be PD. Then

    EL = 100*.6*.03 = Rs.1.80.

    When a company starts to default and the market learns about its distressed state, then the value of plant of machinery falls and LGD rises. Generally, plant and machinery has little sale value in the market unless the buyer is in the same line of business. It fetches little price. The value that can be expected to be received is that of land and building.

    Probability of Default

    The difficulty generally lies with calculation of PD and the following provides methods of calculating PD.

    Method 1

    Let the entire loan accounts of a bank, consisting of 500 accounts, be divided according to the payment record of the borrowers. Suppose there are five categories, and let the frequency distribution be as under:

    Number of defaults

    0

    1

    2

    4

    More than 5

    Number of accounts

    375

    75

    30

    12

    8

    The expected number of defaulting units is 0*375+1*75+2*30+4*12+5*8 divided the total number of units equal to 375+75+30+12+8 = 500. This is equal to (75+60+48+40)/500 = 223/500 = .446. So on an average, the expected number of defaulting units is less than one account. As the number of defaults is low in relation to the total number of accounts, the probability distribution is that of a Poisson Distribution. So the expected number of units with zero default is e (-.446) x(.446)0 / 0! and so on. If we map this with the total outstanding in the accounts, then we can get the expected default amount in each category. The problem arises if the amount involved in the few accounts that default more, is large. In recent years we have witnessed in India large borrowers defaulting in their payments.

    Method 2

    This is based on the options approach and has been elucidated in Galati (2003). Consider an entrepreneur starting an enterprise with equity capital equal to E. Let the enterprise borrow an amount of loan equal to B. Then the total assets that can be created E+ B. Consider the pay-off Figure 1for the borrower. It resembles the pay-off function of a long call where E is the premium paid in the form of equity capital. If the value of the company exceeds B+E, then the company can expect some return on equity. If it falls significantly below B, then the entrepreneur loses E, which is the premium paid, or the equity capital. This is the concept of limited liability. If we now write the option premium equation of Black and Scholes, then

    E = f(B, A, t, r, σ) (1)

    Where E is the option price, B is the strike price, A is the market price or the value of the asset, t is the residual tenure of the loan, r is the rate of interest and σ is volatility. If the company is a listed company, or if we take the share price movement of a company involved in similar business, then σ can stand for volatility of returns from the share. The above equation (1) would then generate a value of A. E is the value of equity that is put upfront by the promoters.

    Once we get a value of A and σ, assuming the default distribution is normal, we can take

    (A - 2σ) to be the lower bound. If (A - 2σ) is less than B, then we have a situation of default. On the basis of this we could calculate the probability of default as the area under the density function between B and (A - 2σ). An empirical probability of default can be obtained from

    [Number of borrowers that defaulted within a year with asset values of (A - 2σ) < B] / [Total number of borrowers with asset values of (A - 2σ) < B in the year]

    Method 3

    The third method uses the CreditMetrics approach of J P Morgan. The examples shown in Figure 2 and Figure 3, taken from Gallati (2003), is for a loan which is rated BBB, the exposure to the loan being $100 million, 6% is the borrowing rate and the tenure being 5 years. Figure 1 gives the transition probabilities of a BBB loan being upgraded, remaining in the same grade, or being downgraded in one year’s time. Loans of different credit ratings enjoy different rates of discount in the market. Better the credit grade, lower is the discount rate. Thus the maximum present value of a BBB $100 million loan, if it gets upgraded to AAA, is $109.37 million. If it remains in the same grade, the value would be $107.55 million. If it moves to the default category, the value would fall to $51.13 million.

    Figure 2 gives the diagrammatic form of the data from Figure 2. We can check that there is a 6.77 per cent probability that the loan value will fall below $102.02 million and a 1.47 per cent probability that the loan value will fall below $98.10 million. From the data it emerges that the expected (mean) value of the loan is $107.09 million and the standard deviation is $2.99 million. Thus, using the normal distribution, the 5% lower limit is 1.65*2.99 = $4.93 million and 1% lower limit is 2.33*2.99 = $6.97 million. These would be extent of the fall in the value of the $100 million loan in case of a downgrade. These values would come close to the values derived from the table of actual default percentage.

    In this third method, we have used the transition probabilities published by J P Morgan. If banks use the same credit rating and transition probability structure that is published, then it would be easy to assign a default probability to each asset category and calculate expected loss.

    Figure 2


    Source: Gallati (2003)

    Figure 3

    Source: Gallati (2003)

  11. Value at Risk (VaR)
  12. This concept is generally used in estimating market risk. We explain this concept with the help of the following example.

    Let us consider stock prices of Bharat Forge, a mid-cap company listed in the National Stock Exchange and the Bombay Stock Exchange, for the last 200 days. If we calculate the daily stock returns, and then construct a frequency distribution of the relative frequency, the result is shown in Figure 4.

    Figure 4: Relative Frequency Distribution of daily returns of Bharat Forge

    Source: Author’s own construction

    Such a diagram can be drawn for any company, and the reader can check that the distributions would look almost like a normal distribution. Only the mean and variance would change. Since this a relative frequency distribution, the long term approximation of which is a probability density function, the sum of the bars would be equal to one. If we put our finger on the horizontal axis and move the finger from right to left, up to the range (minus 2 to minus 1), 95% of the area is covered. Only 5% remains to the left. This implies that with 95% probability we can say that there would be a loss of around 2 to 3 percent on any given day. Or in other words, in the next hundred trading days, on five days, an exposure of Rs.1000 can lead to a loss of Rs.20 or more. This is defined as Value at Risk (VaR), and in the example this is 95% VaR.

    As we have observed, stock returns tend to be normally distributed, and number of times the standard deviation would give a fairly good idea about the VaR with the required probability. Standard deviations are usually quoted on an annual basis. If we assume that there are 254 trading days in a year, then to convert annual standard deviation into 1 day standard deviation would involve multiplying the standard deviation (sd) by square root (1/254) = sd*1/16. Hence,

    Factor Push VaR = Sum of All Single Factor VaR

    Single Factor VaR = Exposure * Maximum Likely Adverse Move

    Maximum Likely Adverse Move = Number of standard deviations * standard deviation * time horizon.

    If we are calculating daily standard deviation, then for daily VaR, time horizon is equal to one. In this case, portfolio VaR is the sum of all individual VaRs. However, in the delta normal method, the correlation between the asset returns become important. For example, if there are two assets, then

    VaR portfolio = Square Root (VaR12 + VaR22 + 2ρ12VaR1VaR2) where ρ is the correlation between the two asset returns.

    This approach can be used to measure market risk. One can also modify this approach and migrate to a Delta Gamma VaR. Here, both the first derivatives and the second derivatives come into play. For bonds, this would mean a combination of Duration and Convexity. For options positions, it would be Delta and Gamma along with other Greeks.

  13. Risk Adjusted Performance and RAROC
    1. Risk Adjusted Performance

    Data:

    Total Loan – Rs.100 crore (Rs.10000 lakh)

    Borrower rated BB – PD or Probability of Default is .05%

    Profit margin of 6%, i.e. 600 basis points

    Loss given default 20%

    Capital Charge fixed by Treasury – Rs.2crore

    Overhead allocation – Rs.50 lakh

    Calculation

    Expected Loss = 10000*.05*.20 = Rs.100 lakh

    Total Revenue = 10000*.06 = Rs.600 lakh

    Risk Adjusted Return (RAR) = 600 – 100 = Rs.500 lakh

    Economic Revenue = RAR – Capital Charge = 500 – 200 = Rs.300 lakh

    Economic Profit = 300 – 50 = Rs.250 lakh

    1. RAROC (Risk adjusted return on capital)

    Income 6.1% - Expected Loss 0.50%

    = Risk Adjusted Income 5.6% - Costs 3.40%

    = Risk adjusted Net Income 2.20% - Tax 0.45%

    = Risk Adjusted After Tax Income 1.75%

    If total assets created are Rs.100000, then Risk Adjusted Net Income = Rs.1750

    Total Risk Capital = Credit Risk Capital 4.40% + Market Risk Capital 1.60% + Operational Risk Capital 2.00% = 8%

    With total assets created Rs.100000, capital required is Rs.8000. If cost of capital is taken as 18%, then capital charge is Rs.1440 = 8000*.18.

    Economic Value Added (EVA) = Risk Adjusted Net Income 1750 – Capital Charge 1440 = Rs.310

    RAROC = Risk Adjusted Net Income / Risk Capital = 1750/8000 = .22 or 22%.

  14. Risk Based Supervision and Risk Management Techniques – A Synthesis
  15. Chart 1 provides an overview of the need for risk management by banks in today’s environment. Banks are in the business of taking risk as they accept deposits from the public and deploy them in financial assets for returns. While deposits have to be returned to depositors on demand, returns from assets are uncertain. Customers have become demanding and they want higher returns. With technological change, newer products with different risk profile are available in the market. Competitive pressure has forced banks to adapt to this dynamic environment and pressure on spreads has made them move to riskier instruments. The need for adoption of risk management techniques has increased and new accounting norms and accounting principles has led to the evolution of risk based organizational structure. Growth in volumes has opened up opportunities for fraud. Technology has opened up creation of complex financial instruments which few people understand. Thus, need for controls have gone up and reporting and MIS has become crucial for decision taking.

    Chart 1

    Source: Author’s own construction

    Advances in technology, ability of money to move across the globe at the speed of light, innovative products, interrelationships across markets, has led to the evolution of risk based management. Banks have moved from traditional supervision (TS) to risk based supervision (RBS). TS is quantifying problems and minimizing risks in individual institutions. RBS is risk based management and recognition of systematic risks across the banking system. As a result we have observed two changes – Basel Accord and CAMELS.

    The core Basel principles are effective supervision, licensing structure, prudential norms, methods of supervision, information requirements and adoption of technology. The full form of CAMELS includes capital adequacy, asset quality, management, earnings appraisal, liquidity appraisal, systems and controls – corporate governance.

    The broad parameters of RBS include a risk based organizational structure, a comprehensive risk management approach, risk management policies, setting of prudential limits, strong MIS, procedures for effective control, separation of risk management from operations and periodical review. A specimen organizational structure is shown in Chart 2.

    Chart 2

    Source: Author’s own construction

    The various dimensions of risk management would include

    1. identifying the risks – credit risk, liquidity risk , market risk , operational risk , interest rate risk, forex risk;
    2. measuring risk – size, duration, probability of adverse consequences;
    3. controlling risks – limits, mitigation, offsetting;
    4. monitoring changes in risks and controls;

    Management of credit risk would involve

    1. setting prudential limits;
    2. credit rating/scoring;
    3. estimation of loan losses;
    4. risk pricing;
    5. portfolio management;
    6. loan review mechanism;
    7. having a credit rating framework

    While acquiring various financial assets, the following would be the scope of credit risk analysis of each of asset.

    1. sector risk – intensity of competition, industry returns, cyclicality of earnings, government policy, regulatory changes;
    2. business risk – relative position;
    3. financial risk – operating margin, Returns on Capital Employed (ROCE), interest cover, liquidity ratios, future cash flows, DSCR, leverage;
    4. management risk – track record, quality of personnel, payment record, financial accounting practices, credibility, support from group companies;
    5. project risk – size of the project relative to net worth, probability & extent of overrun, implementation risk, technology risk, market risk, statutory clearances, financial closure

    In managing market risk, banks need to focus on

    1. liquidity risk – funding risk, time risk, call risk;
    2. interest rate risk – gap risk, basis risk, option risk, yield curve risk, price risk, reinvestment risk;
    3. forex risk;
    4. commodity price risk;
    5. equity price risk

    For control on operational risk, banks need to have in place

    1. processes and systems;
    2. clear delegation of power;
    3. supervision of action at all levels;
    4. a strong human resource base;
    5. awareness of legal provisions of the country of operation;
    6. state of the art technology;
    7. a strong Management Information System;
    8. awareness of regulatory requirements;
    9. assessment of business capacity
  16. Prudential Norms: Income Recognition, Asset Classification, Provisioning and Capital Adequacy
  17. The Reserve Bank of India (RBI) guidelines on the above are available in detail in various master circulars of RBI issued from time to time (see www.rbi.org). These circulars incorporate the Basel guidelines and the changes that have taken. As the master circulars are easily avaiable, there is no need to repeat them here. However, the logic behind the concepts and also the transition from Basel I to Basel II and Basel III needs explanation.

    As per the income recognition norms, income is to be reckoned if income is received. Accrual/accrued income is not treated as income. Once this in place, asset classification becomes easy. If income is received on time, the asset is a standard asset. If not, then according to the delay in receipt of payment, that is principal and/or interest, assets are to be classified as sub-standard, doubtful or loss assets. The latter three comprise non-performing assets (NPAs).

    The question that arises at this stage is, so what if asset classification is required? This is where provisioning comes in. For every NPA created, there is a penalty in the form of provisioning, and this has to come out of after tax income of banks. Profit can be declared, after setting aside the provisioning charges. Higher the NPAs, lower is the declared profits of the bank.

    Eventually, if the surplus income is not high enough to cover provisions, the latter has to be carved out of net worth, and the equity capital of the bank drops. This affects the last prudential norm, capital adequacy. Capital adequacy, defined as the ratio of capital and risk weighted assets restricts the asset creating capacity of banks, by fixing the ratio at 9% currently. Understand the dilemma of the banks. Deposits flow in regularly and they have to be converted into assets. If the capital adequacy norm is not met and the level of capital is not enough to cover for the NPAs, then these deposits have to be channeled to zero risk assets like cash or central government securities. This reduces the yield and hurts profitability of the bank. NPAs lead to provisioning, and this affects capital adequacy and restrict opportunities for exploring high yield assets by banks. One way out is to access the capital market for equity. With deteriorating financial health, it becomes increasingly difficult for banks to raise equity capital from the market. In the case of nationalized banks in India, the onus falls on the central government to infuse capital in these banks and restore capital adequacy.

    The above was defined by the Basel I guidelines and incorporated in the Narasimhan Committee Report. The other elements included inducting professionals in the Board of the banks to improve governance. Basel II first introduced the concept of economic capital, as opposed to regulatory capital laid down by the central bank. The three pillars of Basel II consisted of Minimum Capital Requirements, Supervisory Review Process and Market Discipline. Recognition of credit risk, market risk and operational risk was explicitly made and suggestions were given on how to derive economic capital. Banks were advised to derive their own figures of capital by assessing the risk and calculating the cover for such risks. The interested reader can refer to the BIS Working Papers for detailed investigation.

    Basel III guidelines were a fall out of the sub-prime crisis that generated financial instability all over the world. The basic differences between the Basel II and Basel III guidelines are given in Table 3. A quick look at the two standards will reveal that besides enhancement in capital requirements and disclosure, there are stipulations regarding liquidity requirements and risk management along with capital planning. During sub-prime crisis period, with special reference to Lehman Brothers, it was observed that the company had assets on its books, but there were no takers of these assets. The company was also unwilling to part with the assets at throwaway prices. Further, the company was unable to raise short term funds from the market against these assets. There was, at the end, a liquidity crisis which led to its bankruptcy. The Basel Accord recognized this additional source of risk and included provisions for liquidity enhancement. Besides, capital adequacy, Basel III has emphasized on capital planning and in the following we provide some details.

    Table 3

    Basel II Capital Standards

    Basel III Capital Standards

    Pillar 1: Minimum Capital Requirements

    Pillar 1: Enhanced Minimum Capital and Liquidity Requirements

    Pillar 2: Supervisory Review Process

    Pillar 2: Enhanced Supervisory Review Process for Firm-wide Risk Management and Capital Planning

    Pillar 3: Disclosure and Market Discipline

    Pillar 3: Enhanced Risk Disclosure and Market Discipline

    In addition to higher capital requirements, the new elements in Basel III are New Capital Conservation Buffer, Countercyclical Capital Buffer, Liquidity Standard and Leverage Ratio. These have been stipulated to help financial institutions in times of economic crisis. The idea is to shore up the capital base, beyond capital adequacy standards, such that they can be used as a cushion in times of stress.

    Liquidity Standard includes

    1. Liquidity Coverage Ratio (LCR): to ensure that sufficient high quality liquid resources are available for one month survival in case of a stress scenario.

    2. Net Stable Funding Ratio (NSFR): to promote resiliency over longer-term time horizons by creating additional incentives for banks to fund their activities with more stable sources of funding on an ongoing structural basis.

    3. Additional liquidity monitoring metrics focused on maturity mismatch, concentration of funding and available unencumbered assets.

    With enhanced capital requirements and leverage ratio, the total capital requirements of the banks have gone up.

  18. Capital charge for operational risk
  19. Operational risk has been defined by the Basel Committee on Banking Supervision as the risk of loss resulting from inadequate or failed internal processes, people and systems or from external events. These are risks associated with day to day functioning of a bank and cannot be precisely modelled statistically. It is possible to draw a heat map and then focus on the ones that need to be managed actively.

    According to the RBI guidelines, certain changes have taken place in the financial sector environment which has made explicit recognition of operational risk more relevant.

    “Examples of these new and growing risks faced by banks include:

    • Highly Automated Technology - If not properly controlled, the greater use of more highly automated technology has the potential to transform risks from manual processing errors to system failure risks, as greater reliance is placed on integrated systems.
    • Emergence of e-Commerce – Growth of e-commerce brings with it potential risks (e.g. internal and external fraud and system securities issues)
    • Emergence of banks acting as very large volume service providers creates the need for continual maintenance of high-grade internal controls and back-up systems.
    • Outsourcing – growing use of outsourcing arrangements and the participation in clearing and settlement systems can mitigate some risks but can also present significant other risks to banks.
    • Large-scale acquisitions, mergers, de-mergers and consolidations test the viability of new or newly integrated systems.
    • Banks may engage in risk mitigation techniques (e.g. collateral, derivatives, netting arrangements and asset securitisations) to optimise their exposure to market risk and credit risk, but which in turn may produce other forms of risk (eg. legal risk). ”

    While elaborating on the manifestations of this risk, the RBI guidelines state that the Basel Committee has identified the following types of operational risk events that can lead to substantial losses. These are

    • Internal fraud. For example, intentional misreporting of positions, employee theft, and insider trading on an employee’s own account.
    • External fraud. For example, robbery, forgery, cheque frauds, and damage from computer hacking.
    • Employment practices and workplace safety. For example, workers compensation claims, violation of employee health and safety rules, organised labour activities, discrimination claims, and general liability.
    • Clients, products and business practices. For example, fiduciary breaches, misuse of confidential customer information, improper trading activities on the bank’s account, money laundering, and sale of unauthorised products.
    • Damage to physical assets. For example, terrorism, vandalism, earthquakes, fires and floods.
    • Business disruption and system failures. For example, hardware and software failures, telecommunication problems, and utility outages.
    • Execution, delivery and process management. For example: data entry errors, collateral management failures, incomplete legal documentation, and unauthorized access given to client accounts, non-client counterparty misperformance, and vendor disputes

    In view of the above, RBI has identified the various business lines of a bank and has stipulated total capital charge which is calculated as the simple summation of the regulatory capital charges across each of the business lines. The total capital charge may be expressed as:

    KTSA = {Σ1-3 years Max [∑ (GI1-8*β1-8),0]}/3

    Where:

    KTSA = the capital charge under the Standardised Approach

    GI1-8 = annual gross income in a given year, for each business lines

    β1-8 = a fixed percentage, set by the Committee, relating the level of required capital to the level of the gross income for each of the 8 business lines. The values of the β are detailed in Table 4.

    Table 4

    Business Lines

    Indicator

    Beta factors (%)

    Beta values (%)

    Corporate finance

    Gross income

    β1

    18

    Trading and sales

    Gross income

    β2

    18

    Retail banking

    Gross income

    β3

    12

    Commercial banking

    Gross income

    β4

    15

    Payment and settlement

    Gross income

    β5

    18

    Agency services

    Gross income

    β6

    15

    Asset management

    Gross income

    β7

    12

    Retail brokerage

    Gross income

    β8

    12

    Source: www.rbi.org

  20. Credit Derivatives
  21. Traditionally, banks can reduce risk in a loan portfolio either through diversification with exposure limits, or through occasional sale of the loan assets through creation of a special purpose vehicle and issuance of pass through certificates. In either of the two, the risk of the asset and the asset remained on the books of the lending bank. There was no separation mechanism of credit risk from the asset itself. Credit derivatives enable this separation and the first example is that of a Credit Default Swap (CDS). Figure 4 provides the structure of a CDS.

    Figure 4: Credit Default Swap (CDS)

    Source: The J P Morgan Guide to Credit Derivatives

    The structure of the CDS shown indicates that the protection buyer, who is the lender/owner of the loan asset/a bank, enters into an agreement with a protection seller to receive payment from the protection seller in the event of a default on the part of the borrower. So the default risk is covered, while the loan asset remains on the books of the lender. In return for this protection, the lender has to pay a premium to the protection seller. The terms of the contingent payment would be clearly stated and an event of default would be clearly defined. In case of a default and a contingent payment taking place, the asset would be transferred to the protection seller. The value of this payment would be market linked. The premium to be paid looks like a put premium as it is linked to a right to sell. Note, that during the transaction, the borrower has no information of the transaction. So transaction costs are lower.

    The second example is that of a Total Return Swap (TRS) and the nature of the transaction is shown in Figure 5.

    The total return means interest and fees and any change in value of the loan. The principal does not enter the transaction. In the case, the original holder of the asset pays the total return and in exchange receives a payment linked to the Libor. In case there is depreciation in the value of the total return, the TR payer receives payment from the TR receiver. The structure provides for insurance against depreciation in value of the total return.

    The third example is that of Credit Options (CO) and the structure of the transaction is given in Figure 6. This strategy combines the characteristics of both a CDS and a TRS. We have observed that the structure of a CDS is that of a put option. The protection seeker is the long put, and pays a premium to the protection seller, the short put. The CO combines this with the put buyer simultaneously entering into a swap by which it pays Libor + spread in return for a total return from the asset. The put seller can afford to pay this as it receives put premium on the other leg. In case of default in the underlying asset, the put seller pays the settlement amount and the asset get transferred to the put seller.

    Credit Linked Notes (CLN) are interesting instruments for yield enhancement and default protection, and the basic structure is shown in Figure 7.

    As in CDS, the owner of the assets with credit risk, say a bank, buys protection from a SPV for a premium. The SPV in turn issues credit linked notes to investors, the cash from which is invested in AAA securities (IA). In case the underlying assets default, the SPV liquidates the AAA securities (LP) and pays the bank. Whatever is gained from sale of the underlying assets and also what is left after selling the AAA securities, goes to the investor. The investor bears two sets of risk. First, risk due to default by the underlying asset. Second, due to default of the AAA securities. That is why these credit linked notes have high yield, and the SPV can bear it as it also receives premium on sale of protection.

    Figure 8 gives the full structure of a Credit Default Obligation (CDO). This structure builds in the features of the above discussed four different methods of credit risk management and highlighting why it is financially beneficial for all concerned. It also allows for gradation of investors on the basis of their exposure to risk.

    Figure 8: Collateralized Debt Obligations (CDO)

    The structure has various components. First, against a portfolio of assets, the bank (put buyer) seeks protection from two sources. First, it seeks protection from a SPV for a part of the total assets and pays subordinate swap premium. It is subordinate because a large highly rated counterparty offers the maximum protection and receives super premium. The SPV then floats CLNs to three classes of investors namely, senior class, mezzanine and equity and the proceeds are invested in AAA securities as before. If any of the underlying loan assets default, then the securities are liquidated and any shortfall from the value of the underlying loans is made up first by writing off the equity, then the mezzanine notes and then the senior notes. After that, the super senior protection comes in. As they come in after 10% of asset value erosion, the premium amount is lower.

  22. Theoretical Frameworks for Understanding and Estimating the Relationship Between Risk and Capital
  23. In this section, we examine the relation between risk and capital. What emerged was “have capital, take risk”. Risky projects have higher probability of default, and hence, a higher capital support is necessary to cover loss arising out of risk. Again, if risk is not taken, the opportunity of earning higher returns cannot be there. The following 2x2 matrix brings this out quite clearly.

    Chart 3

    RISK RETURN MATRIX

    RISK

    HIGH

    LOW

    RETURN

    HIGH

    Risk taking

    Does not exist

    LOW

    No takers

    Risk aversion

    Source: Author’s own construction

    Any investment has some amount of risk associated with it. If the Risk is higher, the Returns that are expected should also be higher. Thus, a high Return low Risk asset cannot exist as everybody will start acquiring such assets pushing up their prices and lowering their returns. A high Risk low Return asset would have no takers as everyone would be selling these assets and price would fall and Returns would rise. There can be thus two classes of assets as shown in Chart 3: assets with low risk and low return, and assets with high risk and high return.

    We now turn to different theoretical frameworks for understanding the relation between risk and capital.

    1. Risk Taking in Indian Banks – A Simultaneous Equations Approach
    1. Introduction
    2. First, we will consider the interrelationship between four variables namely actual capital, desired capital, actual risk and desired risk. These variables have been considered in Shrieves and Dahl (1992) and on its extensions by Jacques and Nigro, 1997; Aggarwal and Jacques, 1998, 2001; Rime, 2001; Hassan and Hussain, 2004; Heid et al., 2004; Murinde and Yassen, 2004; Godlewski, 2004; Van Roy, 2003, 2005, and Bouri & Ben Hmida (2006). However, we will build a simultaneous equation model with the above mentioned variables. Indian banks have been exposed to the Basel norms for quite some time. With the opening up of the financial sector, many of them accessed the market for equity capital more than once to shore up their Tier I capital to boost capital adequacy. The first exercise is to show whether Indian banks have been judicious in their risk taking and whether their risk taking has been in line with the capital requirements.

      Before we proceed with the model specification, it is important to understand the implication of Capital Adequacy Ratio (CAR) as a single parameter for monitoring banks by the regulator. It is defined as a ratio with some definition of capital funds in the numerator and total risk weighted assets (RWA) in the denominator. Capital funds would include the components Tier I capital and Tier II capital. Tier I capital includes paid-up capital (ordinary shares), statutory reserves, and other disclosed free reserves, Perpetual Non-cumulative Preference Shares (PNCPS) eligible for inclusion as Tier I capital, Innovative Perpetual Debt Instruments (IPDI), and capital reserves representing surplus arising out of sale proceeds of assets. Tier II capital includes undisclosed reserves, revaluation reserves, general provisions and loss reserves, hybrid capital instruments, subordinated debt and investment reserve account.

      As of today, CAR is 9% in India, which implies that banks can have around eleven times its capital as RWA. As banks are naturally positioned for resources, deposits keep flowing into a bank. These have to be converted to assets, with the overall restriction of 9%. The better the quality of assets, lower is RWA, even with a large actual asset base. The quest for banks is thus for quality assets to lower RWA. On the other hand, greater the credit rating of an exposure, lower is the yield and thus profitability. To look at higher returns, banks have to turn to riskier assets, which would increase the denominator. Thus the other thrust area for the banks is to increase the numerator. If banks cannot access the market for equity for various reasons, they raise Tier II bonds to shore up their capital base and protect their capital adequacy.

      CAR and risk are interdependent. It is postulated that greater risk taking ability is determined by CAR i.e., greater risk taking has to be backed by a larger capital base. Again, greater risk taking leading to acquisition of risky assets, can boost revenue and lead to an increase in reserves and hence Tier I capital. An empirical investigation is then performed to examine the relation between capital and risk taking in private sector banks and nationalized banks and their ability to adjust to unexpected risk. The model introduces the concept of a co-efficient of indexation and we derive values of this from our model.

    3. The Model
    4. The structural form of equations that we consider are given by

      RISKt = RISKt-1 + m(CARtact – CARtreg)… (1)

      CARtact = βCARt - 1act + n(RISKt – RISKt - 1) + (1-n)(RISKte – RISKt - 1) + kCRRISKt-1

      = βCARt - 1act + (RISKte – RISKt - 1) + n(RISKt – RISKte ) + kCRRISKt-1 … (2)

      RISKte = RISKt - 1 + b(RISKt - 1 – RISKt - 2) + cVULNERt-1 + d CLSTATE t-1 + eSIZE t-1… (3)

      where

      RISK – Risk Weighted Assets/Total assets

      CARact– Actual Capital Adequacy Ratio

      CARreg- Regulatory Capital Adequacy Ratio

      CRRISK – Credit Risk defined as Provisions/Total Advances

      SIZE – Level of Bank Deposits

      VULNER – Ratio of Deposits to Risk Weighted Assets

      CLSTATE– Ratio of Government Security Holdings to Total Assets

      t–Time

      e – Expected value

      Regulatory capital is expressed through the required capital adequacy ratio as fixed by the central bank. It is different from economic capital, which explains the amount of capital required to absorb a default. Although the ultimate purpose of regulatory capital is absorption of loss in the event of default, it primarily restricts overexpansion of risk weighted assets in relation to the capital of the bank. If actual capital adequacy ratio exceeds the regulatory capital adequacy ratio, then there is a buffer which leads banks to take more risks which gets reflected in the ratio of risk weighted assets to total assets. This follows from the usual risk return trade-off. If there is flexibility in the banks to take risk, they would take risk, and this is the basis of Equation 1. It also reflects the dynamic nature of portfolio allocation and its relation to capital allocation.

      Equation 2 gives an indexation specification of the actual capital adequacy ratio. The equation indicates that actual capital adequacy is fully indexed to expected risk and partly indexed to unexpected risk, n lying between zero and one. Estimation of the indexation co-efficient is of interest by itself, as it would show to what extent Indian banks have been able to estimate unexpected risk and adjust capital accordingly. The results would be even more interesting as we would be estimating this parameter for both private sector banks and public sector banks separately. The equation also includes CRRISK, which denotes the cover for risk that is already there in the portfolio.

      Equation 3 specifies that the expected level of risk that a bank can take is determined by past risk taking by the bank given by a lag structure and also by VULNER and CLSTATE. Greater the value of VULNER, lower would be the expected risk taking by the bank as, with a similar distribution of returns, the impact of probability of default increases. With respect to CLSTATE, the expected sign is negative as greater proportion of funds in Government Securities implies greater liquidity and lower risk. Larger the SIZE, greater is the expected risk taking as the bank is confident of raising capital any time it needs. It has a relation with goodwill and market perception.

      Equations 1, 2 and 3 give a simultaneous equation structure to capital adequacy and risk taking by a bank. There are three equations in three unknowns RISKt, CAPtact and RISKte. The other variables are treated as exogenous for the model. From here we derive two reduced form equations in RISKt and CAPtact and the estimation results are given in Tables 1 to 4 in Section 3.

    5. The Results
    6. The three equations of our model bring out the interdependence between risk and capital, which is the basis of risk management. The data for the study has been taken for the period 2000 to 2007 on twenty-four nationalized banks and sixteen private sector banks. From the structural equations in Section 2, we derived two reduced form equations in CARt and RISKt . We then ran a regression on the pooled cross section time series data and the estimation results are given in Tables 1 to 4.

      Table-1

      Public sector banks. Dependent variable: CARt

      Variables

      coefficient

      t-value

      R-square

      CARt-1

      0.66

      10.463

      0.519

      RISKt-1

      0.332

      0.957

      RISKt-2

      0.219

      0.343

      CRRISKt-1

      0.154

      1.396

      VULt-1

      0.28

      1.619

      CLSTATEt-1

      -0.487

      -1.659

      SIZEt-1

      0.009

      0.581

      Source: Authors’ own construction

      Table-2

      Public sector banks. Dependent variable: Riskt

      Variables

      coefficient

      t-value

      R-square

      CARt-1

      0.628

      0.109

      0.451

      RISKt-1

      0.342

      1.976

      RISKt-2

      1.097

      0.939

      CRRISKt-1

      0.834

      1.12

      VULt-1

      -1.32

      -1.127

      CLSTATEt-1

      -0.65

      -0.328

      SIZEt-1

      -0.0073

      -0.259

      Source: Authors’ own construction

      The R-square for the public sector banks, where CARt is the dependent variable, is .519 and the coefficient of CARt-1 is statistically significant. That is current period’s capital adequacy is dependent on last period’s level. This is not surprising as year to year variations in this balance sheet figure is not very large. The co-efficient of CLSTATEt-1 is marginally significant, but has the right sign. In Table 2, where RISKt is the dependent variable, only RISKt-1 is statistically significant. Size is not a significant determinant of capital adequacy or risk taking behaviour in public sector banks. However, when we calculate the structural form co-efficients from the reduced from co-efficients, the value of “n”, i.e. the indexation co-efficient for public sector banks is around .4. That is, during the period under consideration, public sector banks have been able to adjust partly to unexpected risk. Their systems and procedures for risk assessment and prediction have been relatively slow to adjust to market variations.

      The estimation results for the private sector banks are given in Tables 3 & 4.

      Table-3

      Private sector banks. Dependent variable: CARt

      Variables

      coefficient

      t-value

      R-square

      CARt-1

      0.615

      6.933

      0.512

      RISKt-1

      0.124

      1.843

      RISKt-2

      -0.735

      -1.433

      CRRISKt-1

      0.522

      2.381

      VULt-1

      0.569

      1.89

      CLSTATEt-1

      -0.112

      -1.984

      SIZEt-1

      0.0079

      0.861

      Source: Authors’ own construction

      Table-4

      Private sector banks. Dependent variable: Riskt

      Variables

      coefficient

      t-value

      R-square

      CARt-1

      -0.712

      -0.211

      0.801

      RISKt-1

      0.288

      1.571

      RISKt-2

      -0.484

      -0.091

      CRRISKt-1

      -0.205

      -0.333

      VULt-1

      -0.216

      -3.449

      CLSTATEt-1

      0.228

      1.437

      SIZEt-1

      0.027

      1.037

      Source: Authors’ own construction

      In Table 3, although the R-square is not very high at 0.512, almost all the coefficients are statistically significant. Past period’s capital adequacy and risk taking behaviour has affected current period’s capital adequacy significantly. These banks, whenever they have taken more risk, have also shored up their capital base. The co-efficients of VULNERt-1 and CLSTATEt-1 are also statistically significant indicating the fact that while these banks have gone on deposit mobilization drives, it has also exposed them to more risk and hence they have raised their capital base and also gone for safer investments in G-Secs. In Table 4, the negative sign of the coefficient of VULNERt-1 indicates that higher the ratio of deposits to total risk weighted assets, lower is the risk taking by the banks. This is intuitive as with higher deposits, the bank is more vulnerable to defaults, and hence takes less risk. Again, when we calculate the structural form co-efficients from the reduced from co-efficients, the value of “n”, i.e. the indexation co-efficient for private sector banks is around .7. That is, during the period under consideration, private sector banks have been able to adjust a great deal to unexpected risk. Being new entrants into the banking business, they have been able to introduce systems and procedures for risk assessment and prediction from the very beginning which has allowed them to react fast to market variations.

    7. Concluding Remarks

    The model and the estimated results enable us to distinguish between the performance of public sector banks and private sector banks in their risk taking behaviour and the rate of adjustment in capital requirements. The results show that, while for public sector banks, capital adequacy is better explained, for private sector banks, risk is better explained. The interesting and important result that we obtain is where we find private sector banks are more equipped to handle unexpected risk. That is their systems and procedures are quite adept in handling variations in the market and also in forecasting repayment patterns of borrowers. Their operational structures allow them the flexibility and the technology platform the required speed of adjustment.

    1. Risk Taking Expectations, Role of Capital and their Impact on Returns – a Study of Indian Banks
      1. Introduction
      2. The purpose of the second exercise is to provide a framework for expectation formation of risk taking by banks, incorporating the role of capital, and their impact on returns. Risk taking is an operational matter for commercial banks. They do it on a regular basis while acquiring financial assets, every moment of time, within norms specified by the management. It is a flow activity of a bank. This activity gives rise to a stock of financial assets, calculated annually, and at times monitored quarterly. Capital, on the other hand, is a stock concept and is not an actively managed on a day to day basis. It is calculated yearly, and at times quarterly, to arrive at the capital adequacy of a bank. It is a variable which determines the risk taking ability of a bank.

        In the first exercise we have presented a theoretical structure where capital is a function of risk and vice versa, and other variables affect the relationships through the desired levels of risk and capital. Here, we incorporate an operational standpoint that it is expectation formation regarding risk that drives a bank and besides capital, other variables influence this expectation. Actual risk taking is a daily affair by a bank, and this in turn determines income. Although periodically capital adequacy is calculated to determine the health of a bank, capital is not something that banks monitor on a daily basis. The following model presents a stylized version of this line of thinking.

      3. The Model
      4. We define a pair of equations

        RISKte = RISKt - 1 + b(RISKt - 1 – RISKt - 2) + cCRRISKt-1 + d(CARt-1act – CARt-1reg) +

        eVULNERt-1 + fSIZEt-1 + ut (1)

        ROA or ROE or NIM = k + n(RISKt – RISKt - 1) + (1-n)(RISKte – RISKt - 1) + vt(2)

        where

        RISK – Risk Weighted Assets/Total assets

        CARact– Actual Capital Adequacy Ratio

        CARreg– Regulatory Capital Adequacy Ratio

        CRRISK – Provisions/Total Advances

        SIZE – logarithm of level of Bank Deposits

        VULNER – Ratio of Deposits to Risk Weighted Assets

        t–Time

        e – Expected value

        ROA - Returns on Assets

        ROE- Returns on Equity

        NIM- Net Interest Margin

        u and v - Random disturbance terms with the usual properties

        The explanation for expectation formation specification in equation 1 is as follows. As specified in various guidelines of RBI, the top management of banks needs to arrive at their risk taking appetite in the beginning of a financial year. The corporate plan would lay down the activity based allocation of capital and risk and hence a forecast for the resultant level of income. This would be monitored and reviewed periodically and necessary calibrations made. Equation 1 lays down a basis for expected risk taking of a bank at the beginning of time t, RISKte.

        It states that if the regulator has not shut down a bank or issued a warning that its risk taking is out of line with its stock of capital, then it would take at least that much risk as it did last year, calculated at the end of period 1, RISKt - 1. Further, from the point of view of growth, the bank would build in some incremental risk, (RISKt - 1 – RISKt - 2).

        As per prudential norms specified by RBI, every bank has to make provisions for non-performing assets (NPAs). These NPAs are in turn a result of risk taking appetite of a bank. If a bank has made sufficient provisions for its existing stock of NPAs, then it has adequate capital cover for its stock NPAs. Thus it can take incremental risk. This is incorporated by including CRRISKt-1 in risk expectation formation.

        As having more capital should prompt a bank to take more risk, the variable

        (CARt-1act – CARt-1reg) states that if a bank’s actual capital adequacy at the end of period t – 1 is more than what the regulator has specified, then it is expected that the bank would take more risk in period t.

        Besides the above explanatory variables, the expected level of risk that a bank can take is determined also by VULNER. Greater the value of VULNER, lower would be the expected risk taking by the bank as, with a similar distribution of returns, the impact of probability of default increases. On the other hand, larger the SIZE, greater is the expected risk taking as the bank is confident of raising capital any time it needs. It has a relation with goodwill and market perception.

        Risk taking by banks means allocation of funds between financial assets. This is turn determines the revenue flow and profitability. Equation 2 specifies that that profitability of a bank depends on its existing asset base given by the term k, and on both expected incremental risk, (RISKte – RISKt - 1) and on unexpected risk, (RISKt – RISKt - 1). In the beginning of period t, banks start off with an expected risk level, but eventually may end up taking more or less risk than they desired. Actual risk taking is dependent on day to day movements in the market and on macroeconomic policy changes both at home and abroad. Our model builds in an indexation coefficient “n” to incorporate the effects of expected risk and unexpected risk on profitability of a bank.

        Equations 1 and 2 complete the model. The relationship between return and risk is specified in equation 2, and the interaction between risk and capital is incorporated in formation of expected risk. Thus the level of capital and its difference with regulatory capital forms risk taking expectations. This forms the framework within which a bank, through its day to day activities, engage in asset acquisition. Day to day changes in macroeconomic conditions and policy measures lead banks to take opportunities in asset acquisition as and when they arise. Thus actual risk taking may differ from expected risk taking. Both, taken together, determine the profitability of bank. Substituting equation 1 in equation 2 and rearranging will generate a reduced form equation given by equation 3, which we will estimate.

        ROAt/ROEt/NIMt = Constant + αRISKt + βRISKt - 1 + γ RISKt - 2 + δCRRISKt-1 + η(CARt-1act – CARt-1reg) + ηVULNERt-1 + λSIZEt-1 + εt(3)

        India is a country where both nationalized banks and private sector banks exist side by side. In our study we will estimate the reduced form equations for both sets of banks, separately, on pooled cross section time series data. Our data is on 27 public sector banks and 23 private sector banks over the period 2007-2012. We present results for all three definitions of returns namely ROA, ROE and NIM.

      5. The Results
      6. The results of our regression are summarized in Tables 1 and 2.

        Table 1: Regression coefficients for public sector banks for different specifications of returns

        CONST.

        RISKt

        RISKt-1

        RISKt-2

        (CARtact - CARtreg)

        CRRISKt-1

        VULNERt-1

        SIZEt-1

        R2

        ROA

        .04

        (3.37)

        -.002

        (-.48)

        -.02

        (-1.48)

        .005

        (1.48)

        .045

        (1.57)

        -.103

        (-.93)

        -.01

        (-2.51)

        -.02

        (-2.0)

        .21

        ROE

        .84

        (4.64)

        -.10

        (-1.68)

        -.32

        (-1.92)

        .10

        (2.01)

        .38

        (.81)

        -2.51

        (-1.41)

        -.18

        (-3.20)

        -.05

        (-2.8)

        .26

        NIM

        .03

        (.02)

        -.006

        (-.09)

        -.011

        (-.61)

        .006

        (1.02)

        .11

        (2.05)

        .13

        (.66)

        -.004

        (-.64)

        0

        (.04)

        .08

        Source: Authors’ own construction

        Table 2: Regression coefficients for private sector banks for different specifications of returns

        CONST.

        RISKt

        RISKt-1

        RISKt-2

        (CARtact -

        CARtreg)

        CRRISKt-1

        VULNERt-1

        SIZEt-1

        R2

        ROA

        .03

        (1.90)

        -.005

        (-.82)

        -.02

        (-1.75)

        -.013

        (-2.14)

        -.14

        (-1.41)

        -.34

        (-3.09)

        -.01

        (-2.12)

        .005
        (5.06)

        .36

        ROE

        .73

        (3.59)

        -.09

        (-1.39)

        -.52

        (-3.1)

        -.2

        (-2.43)

        -.44

        (-1.43)

        -4.54

        -(3.55)

        -.20

        (-3.2)

        .05

        (4.32)

        .40

        NIM

        .05

        (2.62)

        -.005

        (-.69)

        -.023

        (-1.66)

        -.007

        (-.77)

        .04

        (1.19)

        .113

        (.75)

        -.01

        (-1.27)

        .001

        (1)

        .06

        Source: Authors’ own construction

        Observe that the regression results with NIM as the dependent variable is very weak for both sets of banks. However, for ROA and ROE, the explanatory power of our equation, given by the value of R2, is much better for private sector banks than public sector banks. Overall, we may say that our equation formation explains risk taking behavior of private sector banks better than public sector banks.

        For both sets of banks, VULNER is statistically significant with the right sign. The results show that as the ratio of Deposits to Risk Weighted Assets rises, banks tend to take lower risk and hence end up with lower returns.

        The way risk is defined in this paper, it is a ratio of two cumulative figures. Thus the contemporaneous relation between risk and return is not statistically significant for both sets of banks. However, lagged values of risk are. In particular, the extent of significance is greater for private sector banks than public sector banks.

        The relationship between risk and return should be positive. However, we observe a negative sign for both sets of banks for the coefficients of lagged risk. This may indicate that greater risk taking by banks in the past has led to lower returns on assets. Or returns have not increased commensurately with creation of risk weighted assets.

        The difference between actual capital adequacy and regulatory capital adequacy is not statistically significant for both sets of banks. This may indicate that Indian banks have not been proactive in exploiting this difference, but has been more inclined to satisfy the regulator about their balance sheet strength.

        The sign of the coefficient of CRRISK is negative for both sets of banks, but is statistically significant only for private sector banks. This is a bit disturbing as greater provisions to total advances should make a bank comfortable in taking incremental risk and hence give better returns. A negative sign may mean that increased non-performing assets leading to higher provisioning has, both, led to lower returns and also lower risk taking by banks.

        It is interesting to observe that SIZE is statistically significant for both sets of banks, but the coefficients are of opposite sign. Greater deposit mobilization has made income generation easier for private banks. Their goodwill has played a role in their confidence in taking risk and income generation. Private sector banks do enjoy a better valuation in the Indian stock market. Public sector banks, on the other hand, do not enjoy good valuations, and there is a pressure of NPAs. The central government needing to pump in funds to shore up the equity base of public sector banks has made matters worse in the eyes of the public at large.

      7. Concluding Remarks

      In this exercise we have provided an alternative specification of the relationship between risk and capital and their impact on returns. We have done the exercise with three different dependent variables representing returns, namely ROA, ROE and NIM. Our model postulates that all the explanatory variables including capital base, size, past risk taking, cover for non-performing assets affect current risk taking expectations. These, along with actual risk taken, influences returns. Our model explains the overall behavior of private sector banks better than public sector banks. The signs of the coefficients of risk are negative, which is opposite to the direct relationship between risk and returns in theory.

      13. Resolution Strategies for Maximizing Value of Non-Performing Assets (NPAs)

      After providing frameworks for understanding the relationship between risk, return and capital, we now turn to ways and means for recovery from NPAs. For this, we examine the various reasons behind an asset turning into a Non-Performing Asset (NPA) and alternative resolution strategies that can be adopted for recovery from such NPAs. We provide a “state resolution mapping” whereby the reasons for an asset turning NPA and the resolution strategy are shown to be linked together. It is shown that reasons like management inefficiency, industry slowdown, technology, capital structure and product failure will, to a large extent, determine the recovery strategy to be adopted by the bank or financial institution.

      1. Introduction
      2. The issue of Non-Performing Assets (NPAs) in the financial sector has been an area of concern for all economies and reduction in NPAs has become synonymous to functional efficiency of financial intermediaries. From the early nineties till date, the regulators in India, under the general recommendations of the Narasimhan Committee Reports (1 & 2), Verma Committee Report, Basle I, II & III guidelines have continuously provided directives addressed at reducing NPAs. A perusal of the Reserve Bank of India (RBI) circulars in this regard will give the reader a comprehensive idea about the extent of detail in which norms and guidelines have been formulated to arrest the growth in NPAs. It started off with introduction of prudential norms and has delved into adoption of a risk based management system. The Indian financial sector has responded well and adopted the directives given, and the overall health has shown considerable improvement.

        Presence of NPAs indicates asset quality of the balance sheet and hence future income generating prospects of a bank or financial institution. This also requires provisioning which has implications with respect to capital adequacy. Declining capital adequacy adversely affects shareholder value and restricts the ability of the bank/institution to access the capital market for additional equity to enhance capital adequacy. If this happens for a large number of financial intermediaries, then, given that there are large interbank transactions, there could be a domino kind of effect. Low capital adequacy will also severely affect the growth prospects of banks and institutions.

        With weak growth outlook and low functional efficiency, the sector as a whole will not be able to perform its role and will adversely affect the savings investment process. Once we realize this, it is evident that a micro problem of a bank translates into a macro problem of the economy. Capital market development takes a back seat and GDP growth rate weakens. The adverse effects of fiscal deficit loom large and a balance of payments crisis also cannot be ruled out. Banking crisis and foreign exchange crisis get interlinked.

        Besides putting in place checks and balances for acquiring good quality assets, it is true that once an asset turns non-performing, if a resolution strategy for recovery of dues is not put in place quickly and efficiently, these assets would deteriorate in value over time and little value would be realized at the end, except may be its scrap value. That is why, asset securitisation has gained popularity among financial sector players. The literature, however, has not specifically discussed about the various resolution strategies that could be put in place for recovery from NPAs, and in particular, in which situation which strategy should be adopted. The purpose of this paper is to indicate the various considerations that one has to bear in mind before zeroing on a resolution strategy. The details of the strategy would follow after that.

      3. Reasons for an asset turning NPA
      4. The various reasons, either singly or jointly, behind an asset turning NPA can be classified as follows

        • Reasons from the economy side
        • Reasons from the industry side
        • Reasons from the borrower’s side
        • Reasons from the banking system side
        • Reasons from the loan structuring side
        • Reasons from the security side – collateral vs cash flow
        • Reasons from the regulatory side

        From the above, it may be surprising to many that only the borrower is not always at fault. At times, systemic faults can also adversely affect the profitability of financial intermediaries. The following discussion will clarify our position.

        • Reasons from the economy side
        1. Political – mindset regarding paradigm, proactive, fiscally responsible (national income accounts)
        2. Economic – growth, distribution, efficient allocation of resources
        3. Social – acceptability, mobility, education
        4. Technological – advances in use of IT
        5. Legal – Enforceability of loan contracts
        6. Environmental – liberalization & globalisation

        If loan contracts are not easily enforceable, there will naturally be a tendency to default. Opening up of the economy can render companies uncompetitive. Lack of adaptation of IT will make data processing difficult and information dissemination will be impossible. Objective analysis of risk would be difficult and appraisal would remain a subjective matter. Similarly, directed programs of lending can be counterproductive.

        • Reasons from the industry side
        1. Global competition
        2. Cyclical downswing
        3. Sunset industry
        4. Frequent changes in regulatory norms
        • Reasons from the borrower’s side
        1. Misconceived project
        2. Poor governance
        3. Product failure
        4. Inefficient management
        5. Diversion of funds
        6. Dormant capital market
        7. Regulatory changes
        • Reasons from the banking system side
        1. Parameters set for their functioning were deficient: incorrect goal perception and identification – lazy banking
        2. Directed banking and lack of freedom to choose products and pricing
        3. Being unexposed to international marketing methods and products, people lacked training and knowledge resources
        4. Ownership and management were not distinguished – composition of Board of Directors
        5. Lack of systems and procedures – audit and inspections
        6. Banks lacked the ability to handle enormous growth in liabilities and assets
        7. Lack of a mechanism of credit information dissemination
        8. Lack of an effective judicial system for recovery from defaulters
        9. Collateral based lending leading to idle assets
        10. Fixing of price and quantum of loans
        11. Lack of an effective IT system and MIS
        • Reasons from the loan structuring side
          1. High debt equity ratio
          2. Timing of raising equity
          3. Discrepancy between the rate of interest charged and the realistic rate of return
          4. Inconsistency between revenue generation and the loan repayment schedule
          5. Lack of binding penal clauses and performance guarantees
        • Reasons from the security side – collateral vs cash flow

        There is a tendency among banks and institutions to depend excessively on collateral for advancing of loans. While this is important, it presumes from the very beginning that the borrower would default and the security would need to be encashed for recovery of the loan. Clearly, this logic is unacceptable. Emphasis should then be on cash generation and a charge on this should be built into the loan contract through some escrow mechanism.

        • Reasons from the regulatory side

        Frequent regulatory changes can turn assets non-performing. Accounting reason like reduction in income recognition norms from 180 days to 90 days could be one such reason. Pollution related issues could be the other reason. Distance between two sugar mills could be a third.

      5. Recovery Strategies
      6. The various resolution strategies for recovery from NPAs include financial restructuring, change in management, one time settlement, merger, sale to an asset reconstruction company, securitisation of receivables and filing of legal suit. The details under each strategy are given in the following.

        1. Financial restructuring
          • Reschedulement of the principal repayment
          • Reduction in the rate of interest
          • Funding of past due interest into loan or instruments (debt or equity or quasi equity)
          • Funding of future interest
          • Waiver of past simple interest, compound interest or liquidated damages
          • Conversion of loan into equity or quasi-equity
          • Reduction in equity
          • Debt write-off
          • Funds infusion by way of equity or debt for project completion
          • Funds infusion for working capital purposes
          • Escrowing of receivables – Trust & Retention Account
        2. Change in management
          • Change in the promoters
          • Induction of professionals
        3. One time settlement (OTS)
          • Full principal with all past interest and future interest with prepayment premium
          • Full principal with all past interest
          • Full principal with part interest
          • Full principal with full or part interest converted to equity or quasi equity instrument
          • Part principal with the remaining part converted to some equity or quasi equity instrument
          • Part principal and remaining part written off
        4. Merger with another company
          • Nature of the industry – sunrise or sunset
          • Synergy issues
          • Valuation
          • Share swap ratio
          • Tax implications
          a. Sale to an Asset Reconstruction Company (ARC)
          • Nature of the industry
          • Valuation – vintage of machinery, land
          • Provisioning made
          • Failed negotiations
          • Multiple bankers

          The mechanism of sale of a loan to an ARC is as follows. In the Indian context, the time the ARC has for executing a strategy is 5 years. The ARC sells Security Receipts (SRs) to the bank, who invests in them. With these proceeds, the ARC buys the loan from the bank. This asset is then restructured and sold down the line to an investor and the SRs are redeemed. Thus a loan in the books of the bank gets replaced by an instrument, which would be redeemed within five years. The SRs would get valued on NPV basis.

        5. Securitisation of receivables
        6. This goes under the names of CLOs (collaterlised loan obligations), MBS (mortgage based securitisation) and ABS (asset based securitisation). The basic concept is to convert a loan into an instrument, which can be independently traded. This has gained popularity for NPA resolution as it allows lumping of many NPAs or NPAs along with Standard Assets into a single pool and tradeable securities are issued on their behalf. This ensures liquidity and early exit option.

          The originator (bank) floats a Special Purpose Vehicle (SPV) and transfers the asset to it. The SPV then issues securitised notes to an investor whose proceeds go to the originator as payment against the assets transferred to the SPV. The principal and interest payments by the underlying asset get deposited with a Trust, which services principal & interest to the investor.

          Securitization allows issuers to lengthen the maturities of their debt, improve risk management and balance sheet performance, and tap a broader class of investors. It provides liquidity, improved asset liability management and better reinvestment options.

        7. Filing of suit and recovery through liquidation proceedings

        This is an option of last resort. It is time consuming and chances of asset value deterioration are very high.

      7. State-Resolution Mapping (SRM)
      8. We now discuss, as a lender, which strategy to apply and when. Although we will be giving specific options, there may be a non-unique answer to the state. But we will argue what is the best strategy available.

        Any individual company can be understood and analyzed in terms of

        1. The overall state of the economy
        2. Nature of the industry to which it belongs
        3. Overall market growth
        4. Competitive position in the market
        5. Vintage of machinery used
        6. Technology in use
        7. Skill set required and gaps, if any
        8. Management quality
        9. Liability structure (Debt-Equity Ratio)
        10. Outstanding liabilities
        11. Asset quality and provisioning made in the books of the lender

        We will take the above characteristics, two at a time, and design alternative scenarios and associated strategy.

        Case 1

        Market growth

        Competitive position

        High

        Low

        High

        Restructuring/Sale to an ARC

        Restructuring/management change

        Low

        One Time Settlement (OTS)

        OTS, liquidation

        For a company whose competitive position is good and also market prospects are bright, clearly, the lender should stay with the company and provide necessary financial restructuring. On the other hand, a company whose overall market growth prospects are poor and its competitive position is also weak, the lender should do an OTS and exit. The terms of OTS we will discuss in the next section. It is perfectly possible that the borrower may not come forward for a negotiation, and in that case the only option would be to file a suit. An OTS is a highly desirable strategy for a company whose current competitive position is quite good, but where the future market growth prospects are dim. Further, for a company, whose market growth is high, but competitive position is weak, there a financial restructuring should be accompanied by a change in management.

        Clearly we can see how the discussions in Sections 2 & 3 can combine to generate a resolution strategy. In the remainder of the section, we will present various other cases in tabular form. It is left to the reader to analyse the strategy recommended.

        Case 2

        Skill set

        Management Quality

        High

        Low

        High

        Restructuring

        Change in management

        Low

        Merger

        OTS

        Case 3

        Management quality

        Outstanding liabilities

        High

        Low

        High

        Merger

        Restructuring/Merger

        Low

        OTS

        Restructuring/ Change in Management

        Case 4

        Debt-equity ratio

        Nature of the industry

        Growing

        Declining

        High

        Restructuring

        Liquidation

        Low

        Merger/Sale to ARC

        OTS

        Case 5

        Growth prospects of the industry

        Overall macroeconomic condition outward looking, competitive

        Growing

        Stagnating

        High

        Restructuring

        OTS

        Low

        OTS

        Liquidation

        Case 6

        Outstanding liabilities

        Asset quality & provisioning

        Substandard,

        Low provisioning

        Doubtful,

        High Provisioning

        High

        OTS

        Liquidation

        Low

        Restructuring

        OTS

      9. Concluding remarks
      10. In conclusion, it is again emphasized that a conceptual distinction has to be made between past NPAs and future NPAs. Past NPAs are stock. The value is known and recovery has to be maximized. For these assets the recovery strategies are detailed above. However, it is obvious that any asset, when acquired, can turn non-performing. It is this risk for which financial intermediaries are compensated. In order to minimize future risk, the strategies should include

        1. Introduction of a risk based management system
        2. Adoption of Corporate governance practises
        3. Human Resources Management
        4. An efficient Management Information System
        5. Proper IT environment
        6. Adoption of credit rating modules and risk based loan pricing
        7. Securitisation
        8. Market research

        References

        1. Aggarwal R. and Jacques K. T., 1998. Assessing the impact of prompt corrective action on bank capital and risk. Federal Reserve Bank of New York Economic Policy Review, October, 23-32.
        2. Aggarwal R. and Jacques K.T., 2001. The impact of FDICIA and prompt corrective action on bank capital and risk: Estimates using a simultaneous equations model. Journal of Banking and Finance, 25, 1139- 1160.
        3. Jacques, K. and Nigro P., 1997. Risk-based capital, portfolio risk, and bank capital: A simultaneous equations approach. Journal of Economics and Business, 49, 533-547.
        4. Hassan M. K. and Hussain M. E., 2004. Basel capital requirements and bank credit risk taking in developing countries. University of New Orleans/Drexel University, LeBow College of Business, Department of Finance, working paper.
        5. Rime B., 2001. Capital requirements and bank behaviour: Empirical evidence for Switzerland. Journal of Banking and Finance, 4, 789-805.
        6. Shrieves, R. E. and Dahl D., 1992. The relationship between risk and capital in commercial banks. Journal of Banking and Finance, 16, 439- 457.
        7. Heid F., Porath D. and Stolz S., 2004. Does capital regulation matter for bank behaviour? Evidence for German savings banks. Discussion paper n° 03, Series 2: Banking and Financial Supervision, ISBN 3–86558–008–4.
        8. Murinde V. and Yaseen H., 2004. The Impact of Basle Accord regulations on bank capital and risk behaviour: 3D Evidence from the Middle East and North Africa (MENA) region. Third International Conference of the Centre for Regulation and Competition (CRC), on “Pro-Poor Regulation & Competition: Issues, Policies and Practices”, Cape Town, 7-9 September.
        9. Van Roy P., 2003. The impact of the 1988 Basel Accord on banks’ capital ratios and credit RISKtaking: an international study. European Center for Advanced Research in Economics and Statistics (ECARES), working paper, December.
        10. Van Roy P., 2005. The impact of the 1988 Basel Accord on banks’ capital ratios and credit RISKtaking: an international study. Forth Annual Conference of the European Economics and Finance Society on "Economic and Financial Issues in an Enlarged Europe", Faculty of Economics, University of Coimbra, Portugal, 19-22 May.
        11. Godlewski C. J., 2004. Capital regulation and credit risk taking: empirical evidence from banks in emerging market economies. LaRGE, Université Robert Schuman, Institut d’Etudes Politiques, working paper, August.
        12. Abdelfettah BOURI and Anis BEN HMIDA, March 2006, Capital and risk taking of banks under regulation: A simultaneous equations approach in the Tunisian context. Proposition pour le sixième congrès international de l’AFFI : Finance d’entreprise et finance de marché :Quelles complémentarités ?
        13. Master Circular- Prudential Norms on Capital Adequacy and Market Discipline, RBI/2014-2015/90 DBOD No. BP.BC. 5/21.06.001/ 2014-15, July 2, 2014
        14. Master Circular - Prudential norms on Income Recognition, Asset Classification and Provisioning pertaining to Advances, RBI No.2014-15/74, DBOD.No.BP.BC.9/21.04.048/2014-15, July 1, 2014
        15. Master Circular – Exposure Norms for Financial Institutions, RBI/2014-15/78, DBOD.FID.FIC. No. 4/01.02.00/2014-15, July 1, 2014
        16. Basel Committee on Banking Supervision International Convergence of Capital Measurement and Capital Standards, A Revised Framework, Updated November 2005
        17. Basel III: A Global Regulatory Framework for More Resilient Banks and Banking Systems, December 2010 (rev June 2011)
        18. Abreu, José Filipe and Gulamhussen, Mohamed Azzim, 2007, The Relationship Between Capital Requirements and Bank Behavior: A Revision in the Light of Basel II, mimeo pp. 1-34.
        19. Berger, Allen E. ,1994, The Relationship Between Capital and Earnings in Banking, The Wharton Financial Institutions Center Working Paper Series, 94-17, pp. 1-30.
        20. Godlewski, Christophe J. , 2004, Capital Regulation and Credit Risk Taking: Empirical Evidence from Emerging Market Economies, mimeo.
        21. Das, Abhiman and Ghosh, Saibal, 2004, The Relationship Between Risk and Capital: Evidence from Indian Public Sector Banks, mimeo, pp. 1-14.
        22. Altunbas, Yener, Santiago Carbo, Edward P.M. Gardener and Philip Molyneux, 2007, Examining the Relationships between Capital, Risk and Efficiency in European Banking, European Financial Management, Vol. 13, No. 1, pp. 49–70.
        23. Baker, Malcolm and Wurgler, H Jeffrey, (2013), Would Stricter Capital Requirements Raise the Cost of Capital? Bank Capital Regulation and the Low Risk Anomaly, mimeo.
        24. Ahn, Choong Yong, (2001), “Financial System Reform and NPLs in South Korea”, Paper presented at the Shizuoka Asia Pacific Forum, Japan, 1-33.
        25. Banerjee, Abhijit, Shawn Cole and Esther Duflo, (2004), “Banking Reform in India”, BREAD Policy Paper No.006, 2-58.
        26. Batra, Sumant, (2003), “Developing the Asian Markets for Non-Performing Assets: Developments in India”, 3rd. Forum on Asian Insolvency Reform, Seoul, Korea, 2-52.
        27. Claessens, Stijn, (1990), “Experiences of Resolution of Banking Crises”, Paper presented at the PBOC/BIS Seminar on Strengthening the Banking System in China, Beijing, 1-17.
        28. Demirguc-Kunt, Asli & Enrica Detragiache, (1997), “The Determinants of Banking Crises: Evidence from Developed and Developing Countries”, Working Paper, The World Bank.
        29. Demirguc-Kunt, Asli & Enrica Detragiache, (1998), “Financial Liberlisation and Financial Fragility, Paper presented at the World Bank Annual Conference on Development Economics, 1-53.
        30. Hanson, James A, (2001), “Indian Banking: Market Liberalisation and the Pressure for Institutional and Market Framework Reform”, Working Paper No.104, Centre for Research on Economic Development and Policy Reform.
        31. Kearns, Allan, (2003), “Corporate Indebtedness and Liquidation in Ireland”, Quarterly Bulletin, 1-15.
        32. Ketkar, Suhas and Dilip Ratha, (2001), “Securitisation of Future Flow receivables: A Useful Tool for Developing Countries”, Finance Development, 1-10.
        33. Mohieldin, Mahmoud and Sahar Nasr, (2003), “Vulnerability of Banking Systems to Crises in Emerging Markets: The Case of Egypt, Morocco and Tunisia”, Paper presented in the ERF 10th Annual Conference.
        34. Montreevat, Sakulrat and Ramkishen S Rajan, (2003), “Financial Crisis, Bank Restructuring, and Foreign Bank Entry: An Analytical Case Study of Thailand”, Asia Pacific Journal of Economics & Business, 53-77.
        35. Mor, Nachiket & Bhavna Sharma, (2003), “Rooting out Non-Performing Assets”, Paper presented in the 5th Annual Conference on Money and Finance in the Indian Economy, Indira Gandhi Institute of Development Research, Goregaon, Mumbai, 1-22.
        36. Ranjan, Rajiv and Sarat Chandra Dhal, (2003), “Non-Performing Loans and Terms of Credit of Public Sector Banks in India: An Empirical Assessment”, Reserve Bank of India Occasional Papers, Vol 24, No. 3, Winter 2003.
        37. Reto Gallati , (2003), Risk Management and Capital Adequacy, McGraw Hill
        38. The J P Morgan Guide to Credit Derivatives
        39. The Lehman Brothers Guide to Exotic Credit Derivatives