A credit fund is marketed on the yield of the loans it has made.
Between the number on the cover and the money an investor keeps sit four deductions. None of them is concealed, and none of them appears in the headline.
The first is arithmetic rather than cost. A fund quotes the yield on capital it has lent; you earn it only on capital that has been called. Across a three-year ramp with a residual undrawn balance, a 10.50% asset yield becomes 9.40% measured against committed capital before anyone has charged anything.
The second is the fee schedule, and specifically the base it applies to. A 1.25% management fee is a different instrument depending on whether it is charged on invested capital, on the whole commitment, or on gross assets including the fund's own borrowings. At one turn of leverage the third of those is exactly twice the first, at the same headline rate.
The third is credit. A loss assumption drawn from a benign vintage and one drawn from a full cycle are not the same assumption, and the difference between them routinely exceeds the entire fee load.
The fourth is tax, and in this asset class it is usually the largest single deduction. Private credit return is interest. It is ordinary income in the year it arises, with no deferral, no preferential rate, and — where the fund is characterised as an investor rather than a trader — no deduction for the fees paid to earn it.
Every figure the instrument reports is produced by the shared Wealth Intelligence engine from the assumptions on screen. Nothing is sampled, fitted, or drawn from fund data.
The model runs one year at a time on your committed capital. Deployment is set first: capital is called linearly over the ramp until the target proportion is invested, and the remainder earns the cash rate you enter. The fund's asset base is deployed equity multiplied by one plus leverage. From that base it earns the stated asset yield; the undrawn balance earns separately. Credit losses are taken next, then interest on the fund's borrowing, then the management fee and fund expenses, then the incentive fee on whatever profit remains above the preferred return. Tax is applied last, to income rather than to cash. What is left is the cash the investor actually holds.
Year one of the default assumption set, on a $10,000,000 commitment: $3,166,667 deployed and $6,833,333 still undrawn. The loans earn $332,500; the undrawn balance earns $273,333; credit losses take $22,167. The management fee is $39,583 and fund expenses $7,917, leaving $536,167 of income.
No incentive fee is due. The preferred return is 7% on the full commitment, or $700,000, and the year's income does not reach it.
Taxable income is $583,667, not $536,167: the fees are added back because they are not deductible on the setting used here. Tax at 45% is $262,650, and the investor ends the year holding $273,517 — a first-year return of 2.74% on committed capital, against a stated asset yield of 10.50%.
Capital is drawn linearly to the deployment target over the ramp period and stays there. The undrawn balance earns the rate you set, which is the single largest reason a yield on committed capital sits below a yield on invested capital. Nothing here models a capital-call schedule, recycling, or a subscription line; the ramp is a straight line because a straight line is honest about what is being assumed.
Fund borrowing is entered as a multiple of investor equity and applied to the deployed base. It multiplies interest income and credit losses in the same proportion, and the cost of the borrowing is taken between them. Leverage therefore raises income measured on committed capital before it costs anything, which is why the decomposition shows it as an addition and then takes the interest expense back off. Where the cost of debt reaches the asset yield, additional leverage subtracts.
The management fee is charged on one of three bases: invested capital, committed capital, or gross assets including the fund's borrowings. The rate is only half of the term. Add one turn of leverage to the default assumption set and move the 1.25% fee from invested capital to gross assets, and the fee load rises by $732,291 over eight years and the real yield falls by 97 basis points — with no change to the headline rate. The incentive fee is applied to income after losses, interest expense and management fees, above a preferred return expressed on committed capital, and is zero in any year that does not clear it. Any feeder or platform layer is charged separately on invested capital.
Loss is the annual default rate multiplied by one minus the recovery, applied to the fund's asset base. Recovery is treated as immediate. This is the input the instrument is most sensitive to and the one a marketing document is least likely to stress, which is why the sensitivity table varies only the default rate and the recovery and holds everything else at your figures.
Income is taxed at the ordinary rate you enter, in the year it arises. The rate is a single figure and is intended to carry federal, state and any surcharge together. Whether fund-level fees reduce that income is a real fork, and the model does not choose for you: where the fund is characterised as an investor rather than a trader, its fees are miscellaneous itemised deductions and none is allowed, so they are added back to taxable income; where the activity is a trade or business, they are deducted. On the default assumption set that fork is worth 56 basis points of real yield.
The share of the yield you mark as accruing rather than paying is taxed as it accrues but does not arrive as cash. The instrument tracks the two separately and reports both a real yield and a net cash yield. Where the tax due exceeds the cash distributed, it says so and quantifies what the investor has to fund from elsewhere. On the junior-capital assumption set, a 13% coupon half of which accrues produces a real yield of 3.61% and a net cash yield of −1.80%.
Every stage — stated, on committed capital, after leverage cost, after fees, after losses, after tax — is the total for that stage across the holding period, divided by the years, divided by committed capital. That is a simple average rather than an internal rate of return, and it is used deliberately: it makes the stages additive, so the decomposition subtracts to the answer without a reconciling item. It does not compound, and it does not reward or penalise the timing of distributions within the period.
Each of these is a simplification made on purpose. Where one of them matters to a particular fund, it matters more than anything the instrument computes.
The default rate is constant across the holding period. Real credit losses arrive in clusters, and a fund that meets its average through one bad year is not the same investment as one that meets it evenly — the incentive fee in particular behaves differently, because a year of losses does not refund the fees paid in the years before it. The sensitivity table exists because the point estimate should not be trusted on its own.
One yield, one default rate, one recovery. There is no ratings distribution, no maturity ladder, no floating-rate reset, no covenant structure and no concentration. A fund whose result depends on three positions is not described by an average.
The four assumption sets are named for the shape of the terms they contain. They are not observations of the market, not averages of anything, and not claims about how any manager charges or performs. Every figure in them is yours to replace.
A defaulted loan recovers its stated proportion in the year it defaults. In practice recovery takes years, costs money to obtain, and arrives after the interest has stopped. The model is therefore optimistic about credit in a way that grows with the default rate.
All income is treated as ordinary and taxed at a single blended rate. That is close to right for most private credit, and wrong for the parts that are not: equity kickers, warrants, original issue discount on a distressed purchase, and any capital gain on a secondary sale are outside it. State apportionment across a multi-state loan book is not modelled.
Cash received is not re-lent and not reinvested elsewhere, and capital is not returned during the period. The instrument answers what the commitment pays, not what a programme of successive vintages compounds to.
Nothing here assesses underwriting, documentation, seniority in practice, or a manager's workout record. Those decide whether your loss assumption is the right one. The instrument only applies the assumption you give it, consistently.
A $10,000,000 commitment, eight years, a 45% ordinary rate, fees not deductible. Two versions of the same private credit fund, against a public credit alternative yielding 5.50% with no fund layer, no ramp and no defaults.
| Private credit senior, unlevered | Through a cycle 1× levered, 5% default | Public credit 5.50%, no fund layer | |
|---|---|---|---|
| Stated asset yield | 10.50% | 10.50% | 5.50% |
| Income on committed capital | 9.40% | 18.13% | 5.50% |
| Interest on fund borrowing | — | $4,655,000 | — |
| Fees over eight years | $1,078,875 | $1,862,000 | — |
| Credit losses | $465,500 | $3,657,500 | — |
| Yield after fees and losses | 7.47% | 5.41% | 5.50% |
| Tax at 45% | $3,139,031 | $2,786,625 | $1,980,000 |
| Cash received net of tax | $2,839,094 | $1,543,875 | $2,420,000 |
| Real yield | 3.55% | 1.93% | 3.02% |
| Share of the stated yield retained | 34% | 18% | 55% |
The private fund is quoted 500 basis points above the public alternative. After the ramp, the fees, a benign 2% default rate at 65% recovery, and the same tax rate on both, 53 basis points of that survive. The public instrument retains 55% of its stated yield because there is nothing between the coupon and the tax; the private fund retains 34%.
The middle column is the same loan book with a loss rate closer to a full cycle, one turn of leverage, and the management fee charged on gross assets. It earns 18.13% on committed capital before costs and delivers 1.93% after them — below the public alternative, on a headline twice its size. The incentive fee is zero in every year, because the fund never clears its preferred return.
Tax is the largest deduction in the first column: $3,139,031 against $1,078,875 of fees. That is a property of the asset class rather than of the manager. The identical book held by a tax-exempt investor returns 7.47%, and the same book with fund fees deductible against ordinary income returns 4.11% rather than 3.55%.
Three provisions do the work in the tax layer. They are cited because the model applies them, not to characterise any particular fund.
The holder of a debt instrument having original issue discount must include in gross income the sum of the daily portions of that discount for each day of the taxable year on which the instrument was held. Recognition follows accrual, not receipt, which is why the instrument reports a net cash yield separately from a real yield. Payment-in-kind interest is taxed on the same principle, and the model treats the two together as accruing income.
Subsection (h) provides that, notwithstanding the two-per-cent floor in subsection (a), no miscellaneous itemised deduction shall be allowed for any taxable year beginning after 31 December 2017. Subsection (c) directs that regulations prohibit the indirect deduction, through pass-through entities, of amounts that would not be allowable if paid directly by an individual. Where a credit fund is characterised as an investor rather than a trader, the effect is that the investor pays the management fee and is taxed on income measured before it — which is the model's default setting, and reversible.
The character of any item of income, gain, loss, deduction or credit in a partner's distributive share is determined as if the item were realised directly from the source from which the partnership realised it. Interest earned by a fund reaches its partners as interest. That is the provision behind the single most consequential fact in this asset class: there is no structure inside a partnership that converts a loan coupon into anything more favourably taxed.
Nothing here is tax advice, and none of it is jurisdiction-specific beyond the United States federal provisions cited. Rates, characterisation and the treatment of a particular vehicle are questions for your own advisers.
Bring the fee schedule, the fee base, the deployment history and the loss assumption you are being asked to accept. Read personally by a senior specialist, with a written reply usually within one business day.
Confidential PPLI ReviewYour information is submitted over an encrypted connection and handled in accordance with our Privacy Policy. PPLI.com does not sell personal information. Any external introduction is made only with your permission.