Asset Liability Management in Banks: ALCO, Gap Analysis and Duration Gap

Asset liability management in banks is the spine of the CAIIB Risk Management elective: almost every numerical the examiner sets on interest rate risk traces back to a gap statement, a duration number or an ALCO decision. The subject asks one question in many disguises — if rates move 100 basis points tomorrow, how much of the bank's net interest income and how much of its net worth is at stake? This article walks through the ALM organisation, the construction of the rate-sensitivity statement, the arithmetic of traditional gap and duration gap, the economic value of equity, and the behavioural assumptions that quietly decide every one of those numbers.

🏛️ The ALM Organisation: Board, ALCO and the Support Group

ALM is a governance structure before it is a spreadsheet. RBI's ALM System framework places ultimate responsibility on the Board of Directors, which approves the ALM policy and the risk tolerance limits within which the bank may run mismatches. Day-to-day authority is delegated to a board-level Risk Management Committee, and execution sits with the Asset Liability Committee (ALCO).

ALCO is the decision-making body. It is chaired by the MD/CEO (or an Executive Director) and typically includes the heads of treasury, credit, retail and corporate banking, planning, IT and the CFO. Its mandate covers deposit pricing, lending rate spreads over the external benchmark, desired balance sheet profile, and the funding mix — all inside board-approved limits. In most banks ALCO meets at least monthly, and more often when rates are volatile.

Below it sits the ALCO Support Group (or ALM desk), which prepares the analytics: the structural liquidity statement, the interest rate sensitivity statement, scenario runs and the duration profile. Because the same committee that owns pricing also owns mismatch, effective asset liability management in banks is where treasury strategy and business strategy are forced to reconcile.

What the ALM policy must contain

• Definition of the ALM organisation, quorum and escalation for limit breaches
• Time-bucket structure and the tolerance limits for cumulative gaps
• Approved methodology for behavioural slotting, plus a validation cycle
• Prescribed rate-shock scenarios for both the earnings and economic value views
• Reporting formats and frequency to ALCO, RMC and the Board

Candidates should read this alongside the wider taxonomy in Risks and Risk Management in Banks, because ALM sits at the intersection of interest rate risk and liquidity risk rather than inside either one.

💡 Exam Tip: ALCO decides, the ALCO Support Group computes, the Board approves limits. Questions that ask "who fixes the deposit rate?" want ALCO; questions that ask "who approves the tolerance limit?" want the Board.

📊 Rate-Sensitive Assets and Liabilities: Building the Gap Statement

The working document of asset liability management in banks is the interest rate sensitivity (IRS) statement, which slots every rupee of the balance sheet into time buckets by the date on which its rate can next be reset — not by when the contract matures. An item is rate sensitive if it reprices, matures or is contractually repayable within the bucket under review.

Typical classification in Indian bank practice:

• Rate-sensitive assets (RSA): advances linked to an external benchmark or MCLR reset date, money at call, reverse repo, floating-rate investments, and fixed-rate assets maturing in the bucket.
• Rate-sensitive liabilities (RSL): term deposits maturing or repricing, borrowings, repos, refinance, and the rate-sensitive portion of savings deposits.
• Non-sensitive: cash, fixed assets, capital and reserves, non-interest-bearing current deposits, and non-performing advances (an NPA does not reprice — it does not earn).

The gap for a bucket is RSA minus RSL; the gap ratio is RSA divided by RSL. A positive (asset-sensitive) gap means net interest income rises when rates rise and falls when they fall; a negative (liability-sensitive) gap is the mirror image. The standard estimate is:

ΔNII = Gap × Δi × (months remaining in the period ÷ 12)

So a bank with a Rs 500 crore one-year negative gap loses roughly Rs 5 crore of NII for a 100 bp rise over a full year. Because the same bucket structure feeds the funding view, examiners often pair this with the structural liquidity statement in banks — same buckets, different question. The IRS statement asks what mismatch costs; the liquidity statement asks whether the bank can pay.

⚠️ Where Gap Analysis Stops Working

Traditional gap analysis is examinable precisely because it is flawed, and the examiner likes candidates who can name the flaws. Four limitations recur:

1. It ignores everything outside the bucket

Gap measures the effect on accrual income over a short horizon. It says nothing about the change in the present value of long-dated assets — a 15-year housing loan book and a 15-year bond book can look identical in a one-year gap and behave completely differently in economic terms.

2. It assumes a parallel shift and uniform pass-through

Real curves steepen and flatten. Basis risk — the repo rate moving without a matching move in the savings rate, or T-bill yields moving without MCLR — is invisible to a gap report that applies one Δi to every line.

3. It ignores embedded options

Depositors withdraw term deposits prematurely when rates rise; borrowers prepay when rates fall. Both options are written by the bank and neither appears in a contractual gap. Managing them usually requires the hedging instruments covered in Derivatives and Risk Management, such as interest rate swaps and FRAs.

4. It is a single-risk lens

A bank's balance sheet also carries currency mismatch and credit migration. The rupee value of a foreign currency book shifts with the exchange rate — see the treatment of types of foreign exchange exposure in BFM — while the capital consumed by the same book is governed by the standardised approach for credit risk. Longer-horizon threats such as climate stress testing for banks sit even further outside the gap framework.

⚠️ Common Mistake: Slotting an asset by maturity instead of next repricing date. A 10-year floating-rate loan resetting quarterly is rate sensitive in the 3-month bucket, not the over-5-years bucket.

📐 Duration Gap and Economic Value of Equity

Duration converts the entire balance sheet into a single interest rate sensitivity number, which is why the economic value perspective is the second pillar of asset liability management in banks. Modified duration measures the percentage change in the value of an instrument for a 1% change in yield.

The duration gap is:

DGap = D_A − (L ÷ A) × D_L

where D_A and D_L are the weighted average durations of assets and liabilities, and L ÷ A is the leverage ratio of liabilities to assets. The change in economic value of equity follows:

ΔEVE ≈ − DGap × A × Δi ÷ (1 + i)

Worked example: assets Rs 10,000 crore with D_A = 3.0 years, liabilities Rs 9,000 crore with D_L = 1.5 years. DGap = 3.0 − (0.9 × 1.5) = 1.65 years. For a 100 bp rise with i = 8%, ΔEVE ≈ −1.65 × 10,000 × 0.01 ÷ 1.08 ≈ −Rs 153 crore. A positive duration gap therefore destroys net worth when rates rise.

Under the Basel IRRBB standard that RBI's framework follows, banks compute ΔEVE and ΔNII under six prescribed shock scenarios — parallel up, parallel down, steepener, flattener, short-rate up and short-rate down — and a bank is treated as an outlier when the worst-case decline in EVE exceeds 15% of Tier 1 capital. The capital definitions behind that ratio are covered in Basel III Buffers, Liquidity Ratios, Leverage Ratio.

🔁 Behavioural Maturity, CASA and Prepayments

Every number above depends on assumptions, and the assumptions are where real asset liability management in banks is won or lost. Contractual dates are unusable for two large blocks of the balance sheet.

CASA deposits have no maturity. Slotting the entire savings and current book into the earliest bucket would show a grotesque negative gap that no bank actually experiences, because a large core portion stays for years. RBI permits banks to slot non-maturity deposits on the basis of a board-approved behavioural study, splitting the book into a volatile portion (earliest buckets) and a core portion (spread over longer buckets), with the study validated periodically and back-tested against actual outflows. Savings deposits also carry partial rate sensitivity, since the rate can be revised but rarely moves one-for-one with policy.

Term deposits carry a premature withdrawal option; a behavioural study estimates the proportion likely to be withdrawn early and the proportion likely to be rolled over at maturity, and slots accordingly.

Advances carry a prepayment option. RBI has progressively restricted foreclosure and pre-payment charges on floating-rate loans to individual and micro and small enterprise borrowers, which removes the friction that once suppressed prepayments — so retail prepayment speeds now respond directly to rate cuts and must be modelled, not assumed away.

Undrawn cash credit and overdraft limits need the same treatment: only the behaviourally expected drawdown, not the sanctioned limit, belongs in the early buckets. The interaction between these assumptions and funding stability is developed in Liquidity Risk Management.

📌 Remember: Behavioural slotting must be board-approved, documented, periodically validated and applied consistently across the IRS and liquidity statements. A bank cannot treat CASA as sticky in the gap report and volatile in the funding report to flatter both.

What goes to the Board

The reporting pack normally carries the bucket-wise IRS statement with cumulative gaps against limits, NII at risk and EVE at risk under the prescribed shocks, the duration profile of assets and liabilities, deviations from behavioural assumptions observed during the period, and every limit breach with the remedial action taken. Frequency is at least quarterly to the Board and monthly to ALCO.

🧠 Practice MCQs: ALM, Gaps and Duration

Want chapter-wise mock tests with 100+ MCQs? Start practising free →

Master asset liability management in banks and a large slice of the elective paper becomes mechanical: identify whether the question is asking about earnings or economic value, slot by repricing date, apply the right formula, and check the behavioural assumption before trusting the number. Work through more practice on the Risk Management elective tag hub, then test yourself against full-length papers in the CAIIB course before exam day.