WARF Matrix (CLO Matrix)
Definition
The WARF matrix, often just the matrix, is the grid in a CLO indenture that replaces a single maximum WARF with a set of permitted combinations. Each cell pairs a maximum weighted average rating factor with a minimum weighted average spread, and in three-dimensional matrices a minimum diversity score as well. The manager elects a case, one cell, and the portfolio must satisfy that cell's limits; the manager may move to another case whenever the pool complies with it. The matrix lets a manager take more credit risk only if the pool is being paid more for it, or is more diversified, which is the trade the rating agency modeled when it rated the notes.
Why it matters
The matrix is where a CLO manager's strategy becomes visible in a compliance document. A reader who knows only the deal's WARF cannot tell whether 2,850 is comfortable or dangerous; the answer depends on which case the manager has elected and how much spread and diversity the pool carries against it. The matrix also explains a pattern that confuses new CLO investors: two deals from the same manager with the same WARF can have different amounts of trading room, because they sit in different cells. For an investor in CLO equity, the elected case and the cushion to each of its limits say more about the manager's next quarter than the WARF alone.
Common misconceptions
- •The matrix is not a way around the WARF test. Every cell is a WARF limit; the matrix only lets a higher limit apply when the pool has more spread or diversity to justify it, as the rating model requires.
- •A higher-WARF case is not automatically a riskier pool. The extra spread or diversity the case demands offsets the extra credit risk in the agency's model. Risk rises when a manager sits at the limit of every axis at once.
- •Managers cannot switch cases at will to cure a breach. A case election requires the portfolio to comply with the new case at the time of the election; moving to a case whose limits the pool already fails is not permitted.
- •The matrix values are not standardized across deals. Each grid is negotiated at closing and approved by the rating agency for that transaction; a published example grid is illustrative, never a market standard.
Technical details
An illustrative two-dimensional matrix
Rows are minimum weighted average spread; columns are maximum WARF. The figures below are illustrative of a broadly syndicated loan CLO, not any specific deal.
Minimum WAS 3.00 percent: maximum WARF 2,700.
Minimum WAS 3.25 percent: maximum WARF 2,800.
Minimum WAS 3.50 percent: maximum WARF 2,900.
Minimum WAS 3.75 percent: maximum WARF 3,000.
Minimum WAS 4.00 percent: maximum WARF 3,100.
Reading it: a pool earning 3.50 percent over the base rate may carry a WARF up to 2,900; if spreads compress and the pool only earns 3.20 percent, the manager must elect the 3.00 percent row and hold WARF under 2,700, which may force sales of B3 loans.
Adding the diversity dimension
Three-dimensional matrices repeat the spread-by-WARF grid at several minimum diversity scores. A more diversified pool (a higher diversity score) earns a higher WARF limit at the same spread.
Illustration: at minimum WAS 3.50 percent, maximum WARF 2,850 with a diversity score of 60, 2,900 with 70, and 2,950 with 80.
The diversity score itself is a Moody's calculation of how many equally sized, independent obligors the pool behaves like, after grouping by industry; the diversity score page walks through it.
Case elections and the maintain-or-improve standard
The elected case is fixed until the manager elects another. An election must be made to a case the portfolio currently satisfies, and the change is reported to the trustee and shown in the next trustee report.
If the pool fails a limit of its current case and no compliant case exists, the collateral quality test is failing: the manager may only trade in ways that maintain or improve the failing measure. No cash is diverted; that is the job of the coverage tests.
Managers often keep a cushion on every axis of the elected case, typically 25 to 75 WARF points and 10 to 25 basis points of spread, so that ordinary rating migration does not force an election.
Where the matrix comes from
Moody's rates CLO notes by modeling the expected loss on each tranche under many default scenarios. The inputs are the pool's WARF (default probability), diversity score (correlation), recovery rate assumptions and excess spread (loss absorption).
At closing, the arranger and manager propose a set of portfolio combinations; Moody's confirms which combinations still support the assigned ratings. Those combinations are the matrix. It is, in effect, the set of portfolios the ratings were issued against.
S&P's equivalent is the CDO Monitor test: a break-even default rate calculation run against the actual portfolio on each measurement date rather than a pre-approved grid. Many US deals carry both tests; European deals typically use Moody's and Fitch.
How to read it as an investor
The trustee report states the elected case and, for each of its limits, the current value, the limit and the cushion. Read the cushion on all axes: a deal with 200 points of WARF room and 5 basis points of spread room is constrained by spread, not by credit quality.
Case changes over consecutive reports show the manager's direction. Moving toward higher WARF and higher spread in a wide market is the strategy working; moving there because spread compression removed the alternatives is a different signal.
For CLO equity funds, matrix pressure shows up as lower reinvestment spreads and, eventually, lower distributions. AltStreet's Eagle Point Credit review carries the fund-level series that reflects it: NAV per share from $13.39 at the end of 2021 to $5.70 at the end of 2025, and a distribution increasingly characterized as return of capital.
Frequently asked questions
- What is the WARF matrix in a CLO?
- A grid of permitted combinations of maximum WARF, minimum weighted average spread and, often, minimum diversity score. Instead of one WARF limit, the indenture lets the manager choose a case: a higher WARF is allowed if the pool earns a higher spread or is more diversified. The grid is agreed with the rating agency at closing.
- Why not just set one WARF limit?
- Because the rating agency's model cares about expected loss, and expected loss depends on more than the rating mix. Extra spread builds excess spread that absorbs losses; extra diversity reduces the chance of correlated defaults. The matrix encodes those offsets so the manager can adapt to the market without a new rating analysis.
- How does a manager choose a matrix case?
- The manager elects the case whose limits the pool satisfies and, among those, the one that leaves the most room to trade. In a tight spread market a manager might move to a lower-WARF case to escape a WAS floor it can no longer meet; in a wide market it might move to a higher-WARF case to buy cheap B3 loans. Elections are notified to the trustee and appear in the trustee report.
- What does a case change tell an investor?
- Direction of travel. A move to a higher-WARF, higher-spread case means the manager is adding lower-rated loans and being paid for it; a move to a lower-WARF case usually means spreads compressed and the manager could not hold the old row. Repeated moves toward the riskiest corner of the grid are worth a question.
- Is the matrix a Moody's or an S&P concept?
- The WARF matrix is Moody's methodology, since WARF is Moody's measure. S&P's approach to the same problem is the CDO Monitor test, which recomputes break-even default rates for the actual portfolio; some indentures carry both, and the manager must satisfy each agency's test.
- Can a matrix have more than three dimensions?
- Yes. Some indentures add a minimum weighted average recovery rate or a weighted average coupon for fixed-rate assets. Others use a WARF modifier, an adjustment to the effective WARF limit based on the recovery rate, in place of an extra grid axis. The definitions section of the indenture is the only reliable source for a given deal.