Weight each loan's rating factor by par, then read the result knowing that the factors climb steeply as credit quality falls.
WARF, the weighted average rating factor, is the par-weighted average of each loan's Moody's rating factor: WARF = Σ (par × rating factor) ÷ Σ par. On an illustrative $250m direct lending portfolio of ten loans, mostly B1 to B3, it comes to 2,957, between B2 and B3. Averaging the ratings instead of the factors gives B2, a factor of 2,720, and understates the portfolio's risk by 8.0 per cent, because the factors rise steeply as credit quality falls.
Worked in full in The Private Credit Investor by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
WARF is one of the core collateral quality tests in collateralised loan obligations, and it has spread into private credit: middle market CLOs, rated feeder notes, asset-based facilities and many limited partners now ask a direct lending manager for it. Most private loans carry no public rating, so the manager maps each borrower's internal risk rating, or a credit estimate from an agency, onto the rating scale first. The arithmetic after that is simple. Reading it correctly is not.
| Rating | Factor | Rating | Factor |
|---|---|---|---|
| Ba1 | 940 | B3 | 3,490 |
| Ba2 | 1,350 | Caa1 | 4,770 |
| Ba3 | 1,766 | Caa2 | 6,500 |
| B1 | 2,220 | Caa3 | 8,070 |
| B2 | 2,720 | Ca | 10,000 |
The steps are not even. B1 to B2 adds 500, B2 to B3 adds 770, B3 to Caa1 adds 1,280. A Caa1 loan carries 2.15 times the factor of a B1 loan, three notches away.
| Loan | Par | Rating | Factor | Par × factor | Weight |
|---|---|---|---|---|---|
| A | 40 | B1 | 2,220 | 88,800 | 16.0% |
| B | 35 | B2 | 2,720 | 95,200 | 14.0% |
| C | 30 | B2 | 2,720 | 81,600 | 12.0% |
| D | 30 | B3 | 3,490 | 104,700 | 12.0% |
| E | 25 | B1 | 2,220 | 55,500 | 10.0% |
| F | 25 | B3 | 3,490 | 87,250 | 10.0% |
| G | 20 | Ba3 | 1,766 | 35,320 | 8.0% |
| H | 20 | B2 | 2,720 | 54,400 | 8.0% |
| I | 15 | Caa1 | 4,770 | 71,550 | 6.0% |
| J | 10 | Caa2 | 6,500 | 65,000 | 4.0% |
| Total | 250 | 739,320 | 100.0% |
Look up the factor for each loan's rating or credit estimate. A loan with no mapping is often assigned a punitive factor by the governing document, so the mapping is worth doing carefully.
WARF = 739,320 ÷ 250 = 2,957
In Excel, with par in B2:B11 and factors in D2:D11: =SUMPRODUCT(B2:B11,D2:D11)/SUM(B2:B11)
2,957 sits between B2 (2,720) and B3 (3,490): 237 above B2 and 533 below B3. The portfolio is a weak B2 on average, not the B2 the rating mix suggests at a glance.
The two Caa loans are 10.0 per cent of par but 18.5 per cent of the par-weighted factor points. That is the convexity at work: a dollar rated Caa2 counts nearly three times a dollar rated B1. Repay loan J at par and WARF falls from 2,957 to 2,810, although only 4.0 per cent of the portfolio has gone.
The same convexity is why the shortcut fails. Number the notches (Ba3 13, B1 14, B2 15 and so on), take the par-weighted average and you get 15.04, which rounds to B2 and a factor of 2,720. Even interpolating between B2 and B3 gives only 2,751. Averaging ratings treats one notch as one notch wherever it falls; WARF does not.
Suppose the portfolio must keep WARF at or below an illustrative limit of 3,100. Headroom is 143 points.
| Event | New WARF | Change |
|---|---|---|
| Loan G downgraded Ba3 to B1 | 2,994 | 36 |
| Loan J downgraded Caa2 to Caa3 | 3,020 | 63 |
| Loan D downgraded B3 to Caa1 | 3,111 | 154 |
| Loan A downgraded B1 to B3 | 3,160 | 203 |
| Every loan down one notch | 3,834 | 876 |
The first row is a trap: G loses a notch and leaves the Ba category, yet WARF rises only 36 points, because the factors are flatter near the top of the scale. Loan D, the larger B3 borrower, slipping to Caa1 costs 154 points and breaches the limit on its own; the smaller B3 loan, F, would not quite do it. Each $1m that migrates from B2 to Caa1 adds 8.2 points, so $17.4m of the $250m is enough to use the whole headroom.
| New loan rating | WARF |
|---|---|
| Ba3 | 2,849 |
| B2 | 2,936 |
| B3 | 3,006 |
| Caa1 | 3,122 |
A manager near the limit can still add a B3 loan, but not a Caa1 one, and buying a Ba3 loan creates 108 points of room. That is how WARF limits shape origination late in a vehicle's life.
Treating WARF as a rating. It is an average of factors that rise roughly exponentially, so it is pulled up by the weakest names, and two portfolios with the same WARF can hold very different tails: one a barbell of Ba3 and Caa loans, the other all B2. Report the Caa share beside WARF, and the number of loans, which is the other half of the risk; how many loans a private credit fund needs covers it. The second mistake is mapping internal ratings generously. A WARF built on optimistic mappings looks compliant until the agency's credit estimates arrive.
Multiply each loan's par by its rating factor, add, and divide by total par. Here that gives 2,957, a weak B2, with 18.5 per cent of the risk points in 10.0 per cent of the par. Watch the B3 names: one notch down to Caa1 adds 1,280 points per dollar of par, and a single one of them can breach the limit. The free workbook for this book, with the lender's model that stresses a single borrower until its covenant breaks, is on the companion page for The Private Credit Investor.
It depends on the mandate, because WARF is a limit set in each vehicle's documents rather than a market standard. As a reading guide, B2 has a factor of 2,720 and B3 3,490, so a direct lending book of single-B borrowers will usually land between them; the illustrative portfolio scores 2,957. What matters more is the headroom to the limit: here 143 points, used up by loan D, the larger B3 loan, falling to Caa1.
Because the factors are convex: B1 to B2 adds 500 points, B3 to Caa1 adds 1,280. Averaging notches gives the illustrative portfolio a B2, factor 2,720, against a true WARF of 2,957, an 8.0 per cent understatement. The shortcut hides exactly the weak names that carry most of the default risk.
Map each borrower to the rating scale first, using a credit estimate from an agency where the vehicle requires one, or the manager's internal rating translated through a documented mapping, then apply the factors as usual. In the illustrative case the mapping decides everything: one loan of $40m mapped two notches lower, B1 to B3, adds 203 points.
This article is one calculation from The Private Credit Investor. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: credit, private credit and distressed debt → · All 1706 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.