Reader question: A bank can have thousands of loans, bonds, derivatives and securities-financing transactions, yet still be dangerously dependent on one company or one economically connected group. How does the Basel large-exposures framework convert all of those positions into one concentration measure and decide whether the bank has exceeded a prudential limit?
The mechanism is a concentration-control algorithm. First, the bank identifies the true counterparty unit, including groups of connected counterparties. Next, it converts each position into a regulatory exposure value, applies eligible credit-risk mitigation where permitted, looks through funds or structured vehicles when required, aggregates the resulting values, and compares the total with Tier 1 capital.
The core invariant is simple:
Large Exposure Ratio = Aggregate Exposure to Counterparty Group / Tier 1 Capital.
Under the current Basel Framework, the general limit is 25% of Tier 1 capital. A more stringent 15% limit applies to a global systemically important bank’s exposure to another G-SIB. An exposure becomes reportable as a “large exposure” when it reaches 10% of Tier 1 capital.
What this page owns — and what it does not
This article owns the computation:
legal counterparties → connected-counterparty graph → exposure values → permitted mitigation and look-through → aggregate single-name exposure → Tier 1 ratio → limit test.
It does not replace Basel leverage-ratio exposure measurement, CVA capital, CCP default-waterfall mechanics, or the Basel output floor. Those pages own different risk constraints. Large exposures specifically control concentration to one counterparty or connected group.
This is mathematical and regulatory education. It is not financial advice, not a credit recommendation, and not a substitute for the applicable local implementation of Basel standards.
Why risk-weighted capital is not enough
A bank can satisfy ordinary capital ratios and still be vulnerable to a single default.
Imagine a bank with €100 billion of diversified assets and strong aggregate capital ratios, but €20 billion of exposure to one corporate group. If that group fails suddenly, the bank may suffer a loss concentration large enough to threaten its survival even if each individual loan carried an apparently reasonable risk weight.
This is why the Basel Committee treats large-exposure risk as a separate problem from ordinary risk-weighted assets.
The question is not:
How risky is this asset class on average?
It is:
How much can this one counterparty failure hurt us relative to the capital available to absorb the loss?
Step 1: determine the capital denominator
The Basel large-exposures framework uses Tier 1 capital as the denominator.
If:
- Tier 1 capital = €12 billion;
- aggregate exposure to Counterparty X = €2.4 billion;
then:
Large Exposure Ratio = 2.4 / 12 = 20%.
That is below the general 25% limit but above the 10% reporting threshold, so it is a large exposure that must be reported under the framework.
The denominator can move even if the exposure does not
Suppose the counterparty exposure remains €2.4 billion but Tier 1 capital falls from €12 billion to €9 billion after losses elsewhere.
The ratio becomes:
2.4 / 9 = 26.67%.
The bank has crossed the general limit without lending the counterparty one extra euro.
This is a crucial diagnostic: concentration ratios can worsen because the numerator grows or because capital shrinks.
Step 2: identify the true counterparty unit
The hardest part is often not arithmetic. It is deciding which legal entities must be treated as one connected group.
Basel uses two broad connection tests:
- control relationship — one counterparty directly or indirectly controls another;
- economic interdependence — financial problems at one counterparty would likely cause funding or repayment difficulties at another.
If either test is satisfied under the framework, exposures can need to be aggregated.
Connected counterparties form a graph problem
Suppose Bank A has exposures to five legal entities:
- Parent P;
- Subsidiary S1;
- Subsidiary S2;
- Supplier Q;
- Independent Company R.
P controls S1 and S2. Q derives nearly all of its revenue from P and would likely fail if P stopped paying.
The economically relevant counterparty cluster can therefore be:
{P, S1, S2, Q}
rather than four separate names.
The large-exposure engine must build a connected-component map before it sums money.
Why legal-entity separation can hide concentration
A naive system may show:
- P = 7% of Tier 1;
- S1 = 6%;
- S2 = 5%;
- Q = 4%.
Each looks individually below the 10% large-exposure threshold.
But if they belong to one connected-counterparty group:
7 + 6 + 5 + 4 = 22%.
The bank has a material single-group concentration.
Step 3: convert each position into an exposure value
Exposure values are not always equal to accounting carrying values.
Different instruments require different regulatory treatment:
- ordinary banking-book assets;
- off-balance-sheet commitments;
- derivatives;
- securities-financing transactions;
- trading-book positions;
- funds and collective investment undertakings;
- securitisation structures;
- CCP exposures.
The large-exposures engine therefore depends on upstream exposure-measurement modules while preserving its own concentration objective.
On-balance-sheet exposures
For a straightforward loan or bond, the starting exposure is broadly the accounting value subject to the framework’s specific measurement rules and permitted adjustments.
If a bank holds €600 million of an unsecured corporate bond, the initial exposure to that issuer is conceptually close to €600 million before eligible mitigation.
Off-balance-sheet exposures
A €1 billion undrawn commitment is not necessarily treated as a full €1 billion exposure.
Basel credit-conversion factors can convert the off-balance-sheet amount into a regulatory exposure equivalent.
Conceptually:
Exposure = Nominal Commitment × Applicable Credit Conversion Factor.
The correct factor depends on the underlying commitment type and current Basel rules.
Derivatives
Derivative exposure is not simply notional.
A derivative book can have:
- positive replacement cost;
- potential future exposure;
- netting effects;
- collateral;
- wrong-way risk;
- counterparty-specific hedges.
The large-exposure calculation uses regulatory derivative exposure measures rather than treating notional as loss exposure.
This is why the page remains distinct from CVA capital: CVA measures valuation loss from credit-spread deterioration, while large exposures bound concentrated default exposure.
Securities-financing transactions
Repos and securities lending can create exposure to counterparties after recognising collateral and netting under the applicable framework.
A transaction that appears economically secured can still leave residual exposure after haircuts, market moves and netting rules.
Step 4: apply eligible credit-risk mitigation
Basel permits specified risk-mitigation techniques that meet eligibility and minimum-requirement tests.
These can include:
- eligible financial collateral;
- guarantees;
- credit derivatives;
- on-balance-sheet netting.
The principle is not merely to subtract collateral mechanically. The mitigation must satisfy the rules.
Substitution can move concentration rather than destroy it
Suppose Bank A has:
- €500 million exposure to Corporate X;
- a qualifying guarantee from Bank G covering €300 million.
The bank may reduce exposure to X by the protected amount, but it creates or increases exposure to G.
Conceptually:
Exposure to X ↓
Exposure to guarantor G ↑
This is a concentration-transfer problem.
A weak system sees “risk reduction.” A correct system asks where the risk went.
Maturity mismatches matter
A hedge that expires before the underlying exposure can provide incomplete protection.
Basel applies conservative recognition conditions for maturity mismatches. A short-lived hedge should not be allowed to erase a long-lived concentration as if both maturities matched perfectly.
Step 5: look through funds and structured vehicles
A bank can own a fund rather than directly owning the fund’s assets.
If the bank treated the fund manager or legal vehicle as the only counterparty, it could miss a large hidden exposure to one underlying obligor.
The Basel framework therefore requires a look-through approach when applicable.
A simple look-through example
Bank A invests €1 billion in Fund F.
The fund owns:
- 40% Corporate X bonds;
- 20% Corporate Y bonds;
- 40% diversified assets.
The bank’s indirect exposure to X is economically:
€1bn × 40% = €400m
subject to the exact regulatory look-through rules.
If the bank already holds €1.8 billion directly against X, the hidden fund exposure can push the aggregate concentration materially higher.
Materiality thresholds and unknown exposures
Basel uses specific materiality and granularity rules for structures. Where underlying exposures cannot be identified sufficiently, conservative treatment can apply.
The important algorithmic lesson is:
unknown underlying exposure is not automatically zero exposure.
Step 6: aggregate exposures to the connected group
After measurement, mitigation and look-through:
Group Exposure = Σ Adjusted Exposure Values across all members and relevant structures.
The engine must ensure that one economic exposure is not counted twice and that mitigation does not disappear from both the original and substituted counterparty sides.
Step 7: test the Basel limits
General case:
Group Exposure / Tier 1 Capital ≤ 25%.
G-SIB-to-G-SIB case:
Group Exposure / Tier 1 Capital ≤ 15%.
Reporting threshold:
Group Exposure / Tier 1 Capital ≥ 10% means the exposure is a large exposure for reporting purposes.
A worked concentration example
Suppose a bank has Tier 1 capital of €10 billion and the following exposure to one connected group:
- loans: €1.4bn;
- bonds: €0.5bn;
- derivative exposure: €0.3bn;
- indirect fund exposure: €0.2bn;
- eligible guarantee reducing exposure by €0.25bn.
Adjusted group exposure:
1.4 + 0.5 + 0.3 + 0.2 − 0.25 = €2.15bn.
Ratio:
2.15 / 10 = 21.5%.
This is below the general 25% limit but above the 10% reporting threshold.
If the bank were a G-SIB and the counterparty group were another G-SIB, 21.5% would exceed the 15% limit.
The exposure limit is not a credit rating
A counterparty can be AAA and still create excessive concentration.
The framework does not say “safe counterparties may be unlimited.”
It says a single default should not be allowed to create a loss concentration large enough to threaten the bank.
Sovereign and special exposures
The Basel framework contains exemptions and special treatments for specified sovereign, central-bank and other exposures. Qualifying CCP clearing exposures also receive specific treatment.
These are rule-based exceptions, not evidence that the economic risk is literally zero.
A production engine must therefore distinguish:
economically risky from regulatorily exempt under this framework.
Inputs and outputs
A large-exposure engine can require:
- Tier 1 capital;
- legal-entity identifiers and ownership relationships;
- economic-interdependence indicators;
- banking-book exposures;
- trading-book positions;
- derivative exposure measures;
- SFT exposure values;
- off-balance-sheet commitments and conversion factors;
- collateral, guarantees and credit derivatives;
- fund and securitisation underlying exposures;
- G-SIB status;
- exemption flags and jurisdictional implementation rules.
Outputs can include:
- connected-counterparty group identifier;
- gross exposure;
- CRM-adjusted exposure;
- substituted exposure to protection providers;
- look-through exposure;
- aggregate group exposure;
- percentage of Tier 1 capital;
- 10% reporting flag;
- 25% or 15% breach flag;
- reconciliation and provenance records.
Evidence polarity: what supports confidence?
Evidence for a correct implementation includes exposure totals that reconcile to source systems, connected groups that match legal ownership and economic-dependency evidence, mitigation that is recognised only where eligibility rules are met, indirect exposures that reconcile to fund holdings, and ratios that reproduce supervisory reports.
Evidence against confidence includes a corporate group fragmented into many legal names, guarantees reducing the original exposure without creating exposure to the guarantor, unidentified fund holdings treated as zero, stale Tier 1 capital, derivative notionals used directly, or a G-SIB-to-G-SIB exposure tested against 25% instead of 15%.
Failure mode: stale connected-counterparty mapping
Corporate structures change through acquisitions, joint ventures, guarantees and dependency relationships.
A mapping that was correct six months ago can become wrong after a takeover or financing restructuring.
Diagnostic: compare the counterparty graph against current ownership filings, legal-entity data and internal credit assessments.
Failure mode: economic interdependence hidden by legal independence
Two firms can have no ownership link but depend on one another for funding, revenue or repayment capacity.
Diagnostic: flag counterparties with extreme revenue concentration, common funding channels, guarantees or repayment dependence.
Failure mode: CRM double benefit
A guarantee can be mistakenly used to reduce the borrower exposure while the guarantor exposure is never added.
Diagnostic: every recognised substitution should create a paired transfer record.
Failure mode: wrong denominator date
Exposure data may be current while Tier 1 capital is stale.
Diagnostic: enforce common valuation/reporting dates or explicit approved timing rules.
Failure mode: missing look-through
A fund position can hide a large exposure to an underlying issuer.
Diagnostic: compare every fund/structured holding with the current look-through policy and data availability.
Failure mode: exemption leakage
An exemption code intended for one sovereign or CCP exposure can be copied to economically different positions.
Diagnostic: require rule citation and scope test for every exempt exposure.
Counterexample: diversification across instruments is not diversification across names
A bank may hold a loan, bond, swap and repo all involving the same corporate group.
Instrument diversity does not reduce single-name concentration if the same group ultimately owes or supports the exposures.
Counterexample: collateral does not always eliminate concentration
If collateral is volatile, ineligible, maturity-mismatched or itself linked to the same counterparty, the residual or wrong-way exposure can remain material.
Counterexample: a 9% exposure can still matter
An exposure below the 10% “large exposure” definition is not automatically harmless. The 10% threshold is a reporting classification, not a guarantee that concentration below it is economically irrelevant.
Counterexample: compliance today does not guarantee compliance tomorrow
Market movements, drawdowns, derivative replacement cost, corporate mergers or a fall in Tier 1 capital can move the ratio rapidly.
Diagnostics: how to test the engine
- grouping test: split one corporate group across five legal entities and confirm aggregation.
- denominator shock: reduce Tier 1 capital without changing exposures and verify the ratio rises.
- G-SIB test: switch both bank and counterparty to G-SIB status and verify the limit changes to 15%.
- guarantee substitution test: reduce borrower exposure and increase guarantor exposure consistently.
- fund look-through test: inject one 40% underlying name and verify indirect exposure aggregation.
- unknown-underlying test: remove fund transparency and verify conservative treatment rather than zero.
- derivative test: ensure regulatory exposure measure, not notional, feeds the large-exposure numerator.
- breach test: create a 25.1% general exposure and require breach escalation.
- reporting test: create a 10.0% exposure and confirm large-exposure reporting status.
- reconciliation test: aggregate all counterparty groups and trace every component to source data.
What would falsify confidence?
Confidence should be withdrawn if the same economic counterparty can be split across legal entities to avoid the limit; if mitigation removes exposure without transferring or extinguishing risk; if indirect fund holdings disappear from the calculation; if the ratio cannot be reproduced from source data and Tier 1 capital; or if the system cannot explain which Basel rule created each exemption or adjustment.
Alternatives and limits
The large-exposures framework controls single-name concentration. It does not directly solve sector concentration, geographic concentration, interest-rate risk, liquidity risk or broad macro stress.
Banks therefore also use stress testing, economic-capital concentration models, industry limits, country limits, wrong-way-risk controls and internal credit limits.
Those tools complement the Basel large-exposure limit rather than replace it.
Verification and update triggers
Revalidate the engine when:
- Tier 1 capital changes materially;
- a counterparty is acquired, merged or reorganised;
- economic-interdependence evidence changes;
- fund holdings or structured exposures change;
- guarantees or credit protection expire;
- Basel LEX chapters are amended;
- G-SIB designations change;
- local implementation rules change;
- a supervisory report fails reconciliation.
Primary and high-quality references
- Basel Committee on Banking Supervision, Basel Framework — Large exposures, current framework page.
- Basel Committee on Banking Supervision, LEX20 — Requirements, including the 25% general limit, 15% G-SIB-to-G-SIB limit and 10% reporting threshold.
- Basel Committee on Banking Supervision, LEX10 — Definitions and application, including connected counterparties.
- Bank for International Settlements / Financial Stability Institute, Large exposures standard — Executive Summary.
- Basel Committee on Banking Supervision, Final standard for measuring and controlling large exposures.
Educational boundary: This article explains public regulatory mathematics for concentration risk. It does not determine the legal large-exposure treatment of any real institution or transaction and does not provide personalized financial advice.