How CLO Overcollateralization and Interest-Coverage Test Algorithms Redirect Cash: Par Haircuts, OC/IC Ratios, Diversion, Cure Mechanics and Reinvestment Limits

Reader question: A CLO can have enough cash to pay every tranche’s scheduled interest today, yet the transaction documents can still divert cash away from junior debt and equity. What mathematical test causes that diversion?

The key machinery is a set of collateral quality and coverage tests. Two of the most important are overcollateralization (OC) and interest-coverage (IC) tests. The test engine converts the underlying loan portfolio into eligible collateral par or interest income, applies contractual haircuts and exclusions, compares those amounts with the debt associated with a particular test, and produces a ratio. If the ratio falls below the indenture’s trigger, the priority-of-payments machinery can redirect cash that would otherwise flow to junior tranches or equity toward more senior note repayment or other cure actions.

The mathematical idea is simple. The difficult part is defining the numerator correctly. A single defaulted asset, CCC bucket excess, principal loss, trading loss or classification error can alter the test even when the CLO’s bank account has not yet run out of cash.

What this page owns — and what it does not

This page owns:

CLO collateral state + indenture test definitions → OC/IC ratios → pass/fail state → diversion/cure instruction.

It deliberately does not replace the broader securitisation-waterfall article, which owns generic priority-of-payment and tranche cash allocation across securitisations. This page is narrower: it owns the test engine that changes the state of a CLO waterfall.

It also does not replace CreditRisk+ or Gaussian-copula portfolio models, which model future credit loss distributions rather than enforce contractual CLO coverage tests.

This is structured-finance mathematics and document logic, not a recommendation to buy or sell any CLO tranche.

Why a CLO needs coverage tests at all

A CLO typically finances a diversified pool of leveraged loans using several debt tranches plus an equity tranche. Senior notes have priority over junior notes and equity.

If collateral deteriorates, continuing to distribute residual cash to junior stakeholders can weaken the protection available to seniors. Coverage tests act as state-dependent control rules: when a trigger is breached, the waterfall can retain or redirect cash to rebuild senior protection.

That creates a feedback loop:

collateral deterioration → lower test ratio → cash diversion → senior debt paydown → potentially higher future OC ratio.

The test is therefore both a measurement and a contractual control mechanism.

The overcollateralization ratio

A simplified OC ratio is:

OC Ratio = Adjusted Collateral Principal / Relevant CLO Debt Principal.

A public 2026 SEC filing describes CLO overcollateralization in this spirit: total collateral principal is compared with the principal balance of the bonds associated with the relevant ratio, with special treatment for defaulted securities and other assets under the deal rules.

If adjusted collateral is 130 million and the debt associated with the test is 100 million:

OC = 130 / 100 = 130%.

If the contractual trigger is 120%, the test passes.

There is not one universal OC ratio

A CLO can have multiple OC tests, each associated with a different tranche level.

For example, the senior-most OC test may compare collateral with Class A debt. A more junior test may compare collateral with Class A + B + C debt.

Thus:

OC_A = Adjusted Collateral / A debt.

OC_C = Adjusted Collateral / (A + B + C debt).

Because the denominator grows as more junior debt is included, junior OC tests typically have different thresholds and can fail before or after another test depending on the structure.

The numerator is a contractual quantity, not simply accounting par

The biggest implementation mistake is to assume:

Adjusted collateral principal = sum of loan face values.

In reality, the indenture can prescribe haircuts, market-value substitutions or exclusions for particular collateral states.

Public SEC disclosures provide a useful example: for defaulted securities, the value used in an OC test may be the lesser of market value or principal multiplied by an assumed recovery rate.

So a 1 million defaulted loan can contribute much less than 1 million to the numerator even before its final recovery is known.

A default haircut example

Suppose a defaulted loan has:

• par = 1,000,000;
• market value = 320,000;
• specified recovery assumption = 40%.

Recovery-adjusted par:

1,000,000 × 40% = 400,000.

If the deal rule uses the lesser of market value and recovery-adjusted par:

OC numerator contribution = min(320,000, 400,000) = 320,000.

One default therefore reduces effective collateral by 680,000 relative to full par under that stylised rule.

Why CCC buckets matter

Many CLOs allow a limited amount of low-rated collateral to receive full par treatment. Once the portfolio exceeds the contractual CCC concentration threshold, the excess can be haircut under the deal’s specified methodology.

This creates a nonlinear test:

one extra downgrade can matter much more when it pushes the portfolio across the bucket threshold.

Two portfolios with the same average rating can therefore produce different OC results if one sits just below and the other just above the contractual concentration limit.

Interest-coverage tests

A simplified IC ratio is:

IC Ratio = Eligible Interest Income / Relevant Interest Expense.

A public SEC CLO disclosure describes IC in similar terms: interest income is divided by interest expense for the relevant senior classes.

If eligible loan interest is 9.0 million and relevant note interest is 6.0 million:

IC = 150%.

Against a 120% trigger, the test passes.

IC and OC measure different failure modes

OC is mainly a collateral par protection test.

IC is mainly a current interest sufficiency test.

A CLO can pass OC but fail IC if loan coupons fall, assets stop paying interest, financing costs rise or interest proceeds are otherwise insufficient.

It can also fail OC while still passing IC if collateral principal has suffered losses but the surviving loans continue to generate strong interest income.

Therefore one ratio cannot substitute for the other.

Floating-rate assets and liabilities create dynamic IC behaviour

Leveraged loans and CLO liabilities are often floating rate. Their coupons can reset against benchmark rates plus contractual spreads.

If asset and liability spreads do not move together, IC can change even without a default.

A test engine therefore needs the actual coupon state for both assets and notes, including:

• benchmark rates;
• floors;
• spreads;
• payment frequencies;
• non-paying assets;
• deferred or capitalised interest where relevant.

Pass/fail state changes the waterfall

When coverage tests pass, the waterfall can proceed through its normal priority order, eventually allowing residual amounts to junior notes and equity.

When a test fails, the indenture can divert cash that would otherwise have moved downward in priority.

Public SEC filings describe this mechanism directly: if OC or IC compliance tests are breached, cash flows can be diverted from mezzanine debt and equity in favour of more senior CLO debt until the breach is cured.

The test engine therefore produces a state variable such as:

OC_B_PASS = false.

The waterfall engine reads that state and follows the alternate branch required by the indenture.

Cure mechanics: why diversion can repair an OC ratio

Suppose:

• adjusted collateral = 120;
• relevant debt = 102;
• OC trigger = 120%.

Current OC:

120 / 102 ≈ 117.65%.

The test fails.

If 5 of otherwise junior cash is diverted to pay down the relevant debt:

new debt = 97.

Ignoring other changes:

new OC = 120 / 97 ≈ 123.71%.

The denominator shrinks, so the ratio can return above the trigger.

This is why diversion is economically meaningful rather than a bookkeeping penalty.

But a cure can fail if collateral keeps deteriorating

Suppose the same 5 of debt paydown occurs, but another default reduces adjusted collateral from 120 to 112.

Then:

OC = 112 / 97 ≈ 115.46%.

The transaction remains in breach.

A waterfall model that assumes one diversion always cures the test is wrong. The test must be recalculated with updated state.

Interest and principal are separate states

CLO indentures commonly distinguish interest proceeds from principal proceeds.

During the reinvestment period, principal collections are often available for reinvestment in new collateral subject to the indenture’s eligibility and quality tests. After the reinvestment period, principal is more likely to be used for sequential note amortisation under the transaction rules.

Public SEC filings explicitly describe this typical distinction.

A test engine must therefore classify cash correctly. Misclassifying a principal recovery as interest can make an IC ratio look healthier than it really is and send cash down the wrong waterfall branch.

Reinvestment can help or hurt future tests

During the reinvestment period, the collateral manager can replace prepaid or sold loans, subject to contractual constraints.

Reinvestment can:

• restore collateral par;
• change weighted-average spread;
• change rating distribution;
• alter maturity profile;
• change CCC bucket usage;
• change future interest income.

A transaction can therefore pass today’s OC test but become more fragile after a reinvestment decision that reduces spread or increases low-rated concentration.

The weak-link problem: test logic depends on classifications upstream

The arithmetic can be perfect while the result is wrong because one asset has the wrong state.

Examples include:

• a defaulted loan still classified as performing;
• a CCC-rated asset not counted in the concentration bucket;
• a sale receivable treated as cash before settlement;
• a principal distribution classified as interest;
• a loan par amount not reduced after a partial write-down;
• a new purchase included before the contractual effective date.

The coverage-test engine is therefore downstream of reference data, trustee accounting and collateral-state classification.

Inputs and outputs

A robust CLO compliance-test engine can require:

• collateral identifier and current par;
• rating and rating-source rules;
• default status;
• market value where required;
• recovery assumption where required;
• CCC and other concentration buckets;
• interest cash-flow state;
• note principal by class;
• note interest due by class;
• test-specific numerator rules;
• test-specific denominator rules;
• contractual trigger levels;
• payment date and determination date;
• reinvestment-period status;
• rounding and cure mechanics.

Outputs can include:

• adjusted collateral par;
• OC ratio by tranche level;
• eligible interest income;
• IC ratio by tranche level;
• pass/fail state;
• distance to trigger;
• cash-diversion amount;
• post-cure ratio;
• asset-level haircut diagnostics;
• scenario sensitivity to defaults, downgrades and recoveries.

Evidence polarity: what supports confidence?

Evidence for the engine includes exact reconciliation to trustee or issuer-reported OC/IC ratios, correct treatment of defaulted and low-rated assets, consistent principal-versus-interest classification, award of full par only to assets entitled to it, and waterfall behaviour that switches exactly when the contractual trigger is crossed.

Evidence against confidence includes a reported pass when the trustee reports a fail, ratios that do not change after a material default, identical ratios before and after a debt paydown, defaulted assets counted at full par despite contrary deal rules, or cash continuing to equity when the governing test should divert it.

Counterexample: cash in the account does not prove the OC test passes

A CLO may have enough current interest cash to make this quarter’s payments while collateral principal has deteriorated enough to fail OC.

Liquidity and overcollateralization are related but different dimensions.

Counterexample: an asset downgrade does not always reduce OC immediately

A downgrade from BBB to BB may have no direct par haircut under a particular deal’s OC methodology. A downgrade that pushes an asset into a deeply distressed or excess CCC bucket can have a larger effect.

Therefore “rating down one notch → OC falls by fixed amount” is not a valid general algorithm.

Counterexample: high IC does not make junior tranches safe

A CLO can report strong current interest coverage while carrying substantial principal impairment that threatens future cash flow.

IC is a point-in-time interest sufficiency measure under the deal rules, not a complete forecast of tranche loss.

Counterexample: test ratios cannot be compared blindly across CLOs

Two CLOs can both report a 120% OC ratio but use different:

• tranche denominators;
• default haircuts;
• CCC thresholds;
• recovery assumptions;
• asset eligibility rules;
• trigger levels.

The ratio label is not enough. The indenture definition is part of the data.

Weak links in implementation

Wrong denominator. Debt classes included in the test do not match the indenture.

Full-par default bug. Defaulted collateral is not haircut.

CCC bucket bug. Excess low-rated collateral is misidentified.

Interest/principal contamination. Cash is routed into the wrong proceeds account.

Determination-date mismatch. Asset states are taken from the payment date rather than contractual test date.

Stale ratings. Downgrades are not loaded before test calculation.

Reinvestment-state error. Principal is assumed reinvestable after the contractual period has ended.

Rounding-before-comparison. A near-threshold test flips pass/fail because displayed ratio rounding is used instead of full precision.

Diagnostics: how to test the engine

• trustee replay: reproduce a published trustee or issuer OC/IC test for a known determination date.
• single-default test: default one asset and verify the correct haircut path.
• market-value/recovery-minimum test: ensure the specified lesser-of rule chooses the correct amount.
• CCC-threshold test: move one asset across the concentration boundary and inspect the nonlinear effect.
• debt-paydown cure test: reduce the denominator and verify OC improves mechanically.
• continued-deterioration test: combine debt paydown with new collateral loss and recalculate rather than assume cure.
• IC shock test: stop interest on one asset and verify IC, but not necessarily OC, changes.
• proceeds-classification test: swap an interest receipt and principal receipt deliberately and require a reconciliation failure.
• boundary-precision test: run ratios immediately above and below the contractual trigger using full precision.
• waterfall-state test: confirm that the pass/fail output actually changes downstream cash routing.

What would falsify confidence?

Confidence should be withdrawn if the engine cannot reproduce published CLO compliance tests; if a default or haircut does not change adjusted collateral as required; if pass/fail state differs from the indenture under the same inputs; if principal and interest proceeds cannot be reconciled; or if downstream cash routing ignores a test breach.

Alternatives and limits

Coverage tests are contractual protection mechanisms, not probabilistic forecasts. A full CLO valuation model additionally needs default timing, recoveries, prepayments, reinvestment, asset spreads, discounting and scenario paths. A portfolio credit model can estimate future loss distributions but cannot replace the indenture’s deterministic test rules.

For public education, a simplified OC/IC engine is useful because it reveals the control logic. For production valuation, the exact indenture language owns the truth.

How this connects to the surrounding knowledge estate

The general securitisation waterfall explains how cash is ranked across tranches; this page supplies one important state signal that can change that ranking. Rating-transition algorithms can generate downgrade scenarios that feed CCC-bucket stress. CreditRisk+ and Gaussian-copula models can generate portfolio default scenarios, but the contractual OC/IC engine determines how those realised states affect the CLO’s cash-control rules.

Verification and update triggers

Preserve the indenture version, test definitions, tranche balances, collateral snapshot, ratings, default states, market values, recovery assumptions, concentration buckets, proceeds classifications, determination date and full-precision ratios. Revalidate after amendments, refinancing/reset transactions, changes to rating-agency methodology embedded in the indenture, collateral-manager substitutions, trustee-system migrations or any mismatch with an official trustee test report.

Primary and high-quality references

• U.S. Securities and Exchange Commission filing, 2026 CLO compliance-test disclosure, including current OC and IC ratios and a description of defaulted-asset treatment.
• U.S. Securities and Exchange Commission filing, CLO structure and waterfall disclosure, describing interest and principal waterfalls and cash diversion after OC/IC breaches.
• U.S. Securities and Exchange Commission filing, 2026 CLO principal-waterfall and reinvestment-period disclosure.
• Bank for International Settlements, BIS research on leveraged loans, securitisation and structured credit.
• U.S. Federal Reserve, Federal Reserve research and financial-stability materials on leveraged lending and CLO markets.

Educational boundary: This article explains CLO contractual coverage-test mathematics. It does not evaluate a particular CLO, recommend a tranche or provide personalized financial advice.