ALM and Liquidity Risk
Scene at a typical US Savings Bank in 1979
Customers walk into the bank to place money in savings accounts - @ 3% Bank gives out housing loan/mortgage - @ 6% Bank makes spread of 3% (without doing anything actually) Satisfied CEO leaves at 3 PM to play golf 3-6-3 banking: Easiest thing in the world!
What makes it so easy
US Federal Reserve follows regulated interest rate policy Note: Banks are borrowing 3-6 m and lending 10-12 yrs Current practice yields profit as long as: 3m rate < 10 yr rate But is there a guarantee that it will?..
Scene at US Fed Reserve: Oct 1979
Fed Reserve Chairman makes major announcement on Fed Policy change Short term interest rates were made floating Result: 3m rates zoom to nearly 20% 10 yr money remains at 6% What happens to the Savings Banks?
Nightmare ending to the story
Savings banks cannot afford to borrow 3m money and incur negative spread of up to 14% Banks forced to sell loans at ridiculous rates to raise cash Result: Entire industry practically wiped out overnight Morale of the story: No regard to integrated planning of assets and liabilities can ruin you.
Direct outcome of this crisis
Creation of a new field of study and practice known as: Asset Liability Management
Defined as the strategic positioning of the balance sheet to: – Immunise profitability from core banking activities – Optimize intrinsic value of the bank – Ensure ability at all times to meet claims by creditors
What ALM is about
Banks make profit by playing the spread between asset/liability interest rates Thus ALM in summary is all about managing two central risks: – Interest Rate Risk – Liquidity Risk For banks with forex operations: – Currency risk (additionally)
Office of Thrift Supervision (OTS), regulatory authority for savings banks (USA) issued Thrift Bulletin -13 (TB-13) in 1989 – Comprehensive guidelines for interest rate risk management – Focused on advanced measures – Gave a boost for ALM/Risk Management technology vendors Reserve Bank of India issued ALM Guidelines in February 1999
Managing risks in general
‘Managing’ a risk has several distinct steps: – Identifying the risk(s) – Measuring the risk(s) – Attempt to restrict the risk involved by: Placing limits on the risk(s) - imposing discipline Constantly monitoring the risk (to avoid unpleasant surprises) Taking suitable measures to avoid breaching the limits
Identifying the Risk(s)
Interest Rate Risk – Arises because a bank has differing ‘repricing’ profiles on assets and liabilities – Dilemma is that banks make money as a result of this mismatch Hence in practice, the idea is to limit the mismatches rather than aim at zero mismatches – Basis risk: When two floating rates are involved
The risk that a bank may not be able to generate cash to meet expected/unexpected claims. – One of the most difficult risks to define and model – Can occur due to: Systemic or non-systemic factors Market shock leading to sudden drying up of activity No lenders/buyers available in the market Bunching up of outflows and counter-party limits become an obstacle to further borrowing. – Bank paying more to borrow in tight situations also qualifies as liquidity risk.
The risk of adverse movements in a particular currency – Position limits combined with sensitivity limits can reduce the potential exposure – Ex. If currency A is more volatile than currency B then smaller position in A has similar risk to larger position in B.
Measuring the Risks
Interest Rate Risk – Measurement tools/techniques Gaps (Static and dynamic) Duration Value-at-Risk Simulation – To perform a diagnosis a doctor performs multiple tests (cannot rely on one technique alone) – Banks adopt these techniques in the above order (as per increasing sophistication levels)
Measuring the Risks
– Measurement tools/techniques
Gaps (static/dynamic) Refined gap - cash flow ladder Stress tests
– Lack of tools/techniques only shows how much of a grey area liquidity risk is.
– Gap reports – Value-at-Risk
Measuring Interest Rate Risk
A gap report can be of two types: – Static Gap – Dynamic Gap It involves taking all interest rate sensitive items and classifying them as per the time period in which the interest rate will be reset (repricing). Static gap assumes the current balance sheet and positions stay as they are and are allowed to run down gradually Dynamic gap assumes that new business will be added on
Consider a 3m fixed deposit (Rs. 100) contracted at 10%. – At the time of maturity the depositor can renew – Deposit thus reprices after 3 months. Consider a 3m CP linked Floating Rate Bond (Rs. 80) – Bond matures in 2002 (3 years) – Interest rate is reset after every 3 months – Bond thus “reprices” in 3 months ‘Repricing Gap’ in 3 months is difference between assets and liabilities (Rs. 20)
Since measuring gaps for each day is too cumbersome, one does it for time intervals (called time buckets) Time buckets should be more refined where bank perceives more activity RBI issued guidelines in 1999 specifying the following time buckets: <1m 1-3m 3-6m 6-12m 1-3y 3-5y >5y Non-sensitive
For fixed rate items (term deposits, investments, bills discounted ) – Classification as per maturity For floating rate items – Classification as per next repricing date – PLR linked assets are classified as per expected date of PLR change Items like cash, current account, capital, funds payable, other assets/liabilities etc. are classified in Non-sensitive category.
300 250 200
Assets Liabilities Gap Cum Gap
<1m 70 135 -65 -65
1-3m 45 110 -65 -130
3-6m 280 250 30 -100
6-12m 200 210 -10 -110
1-3y 200 120 80 -30
3-5y 225 90 135 105
>5y 130 95 35 140
NS 50 190 -140 0
Cumulative Gap after 1yr is Rs. (110). – So what? Gap Figure must be used to arrive at a kind of loss estimate This is called Earnings at Risk (EaR): – It is a loss estimate given the current gap position and an adverse rate movement – Since gap in <1m bucket is negative, bank will incur a loss if interest rates increase.
EaR calculation in <1m bucket: Gap = -65 Say interest rates rise by 1% On average, bank incurs the additional cost over 350 days (360-15 days) EaR = (-65)*(350/365)*(1/100) = Rs. 0.62 This is the figure that banks need to monitor to cap interest rate exposure
Duration Gap Methodology
Problem with traditional gap methodology is that it does not recognize sensitivities of various assets/liabilities Can be resolved using “duration” – Very popular concept in bond markets – Defined generally as sensitivity of bond value to change in interest rates – Bond prices decrease when interest rates increase – Greater the duration; greater the sensitivity Modified duration is the (%) change in price due to a 100 bps parallel shift in the yield curve
A bank has Rs. 100 asset (of Mod. Durn 3) A bank has Rs. 90 liability (of Mod. Durn 2) Difference between assets/liabilities is Equity – used as a buffer to absorb losses Interest rates increase by 100 bps – Loss in assets (3% * 100) = 3 – Gain in liabilities (2% * 90) = 1.80 – Net loss to the Bank = 1.20 (Erosion in capital funds) – Sensitivity of capital = 1.20/10 = 12% – Mod Duration of Equity = (DA * A - DL * L) / (A - L)
Bank is assumed to have issued bonds (liabilities) and invest in bonds (assets) – Deposits/loans can be treated as bonds – These have to be marked to market (using appropriate discount rates) – Duration can be computed easily Duration is an “economic value” risk measure Earnings at Risk is an “earnings” risk measure Both need to be considered
Value at Risk
Defined as the maximum loss in value expected to occur for a given – Time horizon – Confidence level (Probability) Say VaR at 99% confidence over one day for a trading position is Rs. 2 crore – 99 days out of 100 the loss will be less than Rs. 2 crore – On one day the expected loss is not determinable – VaR is now widely used as a trading risk measure
Value at Risk
Does VaR have a place in ALM, which is generally not concerned with trading risks? Refer duration discussion where all assets/liabilities are expressed as bonds – VaR for a bond is expressed as loss in value; not loss in income (bond is fixed income instrument) – Floating rate bonds reprice quickly - hence small loss in value but potentially large loss in income VaR in the ALM context applies to loss in economic value or market value
Value at Risk & Simulation
Suppose a 3m deposit is taken today
– It will be rolled over 3 times over a one year horizon – At what rate will those rollovers happen?
We do not know but if we make thousands of guesses, the average of those guesses is the best estimate available today
This is the idea behind simulation – Make thousands of guesses about how the 3m deposit rate will evolve over one year Each guess is technically known as a “scenario” – Under each scenario the total interest expense can be calculated – The mean of those expenses is the best available “single guess” for the expected deposit expense over one year But how does one guess an interest rate? – Monte Carlo Simulation
The purpose of simulation is to generate interest rate scenarios – Simulation is not a measure of risk (like duration, VaR etc.) – The variability in income as a result of various interest rate scenarios is the risk – Preference is to stay as close to budgeted levels irrespective of the interest rate scenario Only recourse is to hedge using off-balance sheet items (most efficient means) The result of a simulation is a management action on reducing NII variability
The Big Picture
Gaps are simple to use and are quick & dirty Duration needs large amounts of data – Talks about mismatches and sensitivities – Does not talk about probability of losses VaR uses probabilities (very useful) – Large amounts of data required (headache) – Statistical/econometric skills for modelling – What happens beyond 99% probability (nobody knows the extent of the loss) – “Loss in value” risk measure
The Big Picture
Simulation is state-of-the-art – Can handle “loss in income” and “loss in value” measures – Resources - intensive data, qualified staff, huge computational power Expensive proposition ALM is not only about complex calculations – It is about judgements and prudence in taking risks given expectations of market movements – ….and making money as well in the process
ALM– A Case Study
Liquidity position of a bank is as under : ALCO approval for call borrowing Rs.800 Cr Currently borrowing in call Rs. 500 Cr Existing liquid assets : Rs.1000 Cr Further outflows during next 14 days – Retail Deposits 100 Cr – Corporate Deposits 300 Cr Disbursements of loans lined up Rs.300 Cr Additional investment in SLR bonds required in next fortnight Rs.150 Cr Guarantees of Rs.100 Cr likely to devolve
Case Study cont…
Requirement : Higher profits through higher NII Other factors : Capital adequacy at 9.10% Negative Gap to be capped at 20%
Total outflows during the next 14 days : Retail Deposits (assuming 75% renewal) Corp. Deposits Disbursement SLR Gurantee Total 25 Crs 300 Crs 300 Crs 150 Crs 100 Crs 875 Crs
Call Borrowing would increase as under : 500 Cr + 875 Cr 1375 Cr ALCO limit of Rs.800 Cr being breached Capital adequacy ratio likely to fall below 9%
Gap position in next 14 days as under : Inflows 1000 Cr ( liquid assets ) Outflows 1375 Cr Gap (-) 375 Cr Negative Gap as a percentage of outflows : 27% (Higher than the 20% regulatory limit)
Activate all resource raising units Retail to be asked to bring additional Rs.50 Cr Corporate banking to ensure renewal of at least 75% of deposits maturing Explore possibility of renewal of guarantee
Remedial Measures contd…
Sale of liquid assets to bring borrowing within ALCO limits and help maintain CAR Explore possibility of interest rate swaps which provide scope for NII without funding requirements Explore interest earning opportunities by switching securities