The arithmetic of a trade error is simple. The netting policy decides how much the adviser pays, and it is what a reviewer tests.
A trade error reimbursement is the loss each client account suffered from each error, error by error, paid by the adviser, while any gain from an error stays with the account. In an illustrative quarter with four errors across two funds, that comes to $48,650. Netting gains against losses across errors would cut the adviser's cheque to $38,650, which is why the netting policy, not the arithmetic, is what an examiner reads first.
Worked in full in The Private Fund Compliance Officer by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Trade errors are rarer in private funds than in a trading book, but credit funds, hybrid funds and any manager holding listed positions or FX hedges have them every quarter. Each one needs a number, a decision on who bears it and a record that the decision followed the written policy. The calculation below is generic; the netting rules your firm applies must be the ones in its policy and fund documents.
| Error | Fund | What happened | Result for the fund ($) |
|---|---|---|---|
| E1, fat finger | A | Bought 50,000 shares instead of 5,000 at $42.10; the 45,000 excess sold next day at $41.35, commission $0.01 a share each way | −34,650 |
| E2, wrong account | A | Sold $2,000,000 of a loan at 98.50 that Fund B should have sold; bought back at 98.25 | 5,000 |
| E2, wrong account | B | The intended sale went through four days late at 98.25 | −5,000 |
| E3, duplicate hedge | A | A 3,000,000 EUR forward booked twice; the duplicate unwound at 1.0850 against 1.0880 | −9,000 |
| E4, oversale | A | Sold 10,000 shares too many at $18.40; bought back at $17.90 | 5,000 |
Loss on a correcting trade = quantity × (price paid − price received) + correction costs.
E1: 45,000 × (42.10 − 41.35) = 33,750 of price loss, plus 900 of commission on two legs = 34,650. The commissions belong in the loss: the fund would not have paid them without the error.
E3: 3,000,000 × (1.0880 − 1.0850) = 9,000.
E2 is two results, one per fund. Fund A sold high and bought back low: a gain of 2,000,000 × 0.25% = 5,000. Fund B sold the same loan four days late at the lower price: a loss of 5,000. One error, two clients.
Measure in each account separately, before any decision on netting. In a spreadsheet keep one row per error per account, with columns for the trade, the correction, costs and the result; =SUMIF(Result,"<0") then gives the gross losses.
| Policy | Adviser pays ($) | Comment |
|---|---|---|
| No netting: each loss reimbursed, each gain kept by the fund | 48,650 | Fund A keeps 10,000 of gains |
| Net within each error, within each fund | 48,650 | Same here: each error touches each fund once |
| Net within each error, across funds | 43,650 | Fund A's gain on E2 pays Fund B's loss |
| Net across all errors, fund by fund | 38,650 | Fund A nets to −33,650, Fund B to −5,000 |
The no-netting answer is the cleanest to defend: the adviser makes each account whole for each mistake and never uses one client's windfall to pay for another error. The third row is the one to avoid in every case, because it uses Fund A's money to make Fund B whole; Fund A and Fund B are different clients with different investors. The fourth row, netting across errors within one fund, is the contested one. It saves the adviser $10,000, or 20.6% of the bill, and LPs and regulators look hard at it. It is defensible only if the policy says so and the fund's documents disclose it, and many advisers do not allow it at all.
If reimbursement is paid 30 days after the error is corrected and the policy compensates for the delay at 5% a year: 48,650 × 5% × 30 ÷ 365 = 199.93, so the adviser pays 48,850 in all.
Small, but leaving it out tells the reviewer that the firm treats reimbursement as a favour rather than an obligation. The same interest logic applies to fee refunds, worked in how to calculate a fee overcharge refund with interest.
| Gain on E4 ($) | Adviser pays, no netting | Adviser pays, netting across errors | Fund A net result | Adviser saves |
|---|---|---|---|---|
| 5,000 | 48,650 | 38,650 | −33,650 | 10,000 |
| 20,000 | 48,650 | 23,650 | −18,650 | 25,000 |
| 40,000 | 48,650 | 5,000 | 1,350 | 43,650 |
| 60,000 | 48,650 | 5,000 | 21,350 | 43,650 |
With one lucky error, netting erases the adviser's liability to Fund A entirely. Under a no-netting policy, the adviser's cost does not depend on luck. That asymmetry is the argument to put to a COO who sees the netting policy as a cost saving: the saving comes only in quarters when the market happened to move the right way, and it is exactly the pattern an examiner tests for.
The common error is measuring an error at fund level instead of account by account, which hides cross-client netting inside a single number. E2 shows how: booked as one error, it nets to zero and nobody pays Fund B. The second mistake is leaving correction costs, commissions, FX spreads and break costs, in the fund. Both errors are easy to make in a spreadsheet and easy to find in a sample, the kind of exception discussed in how to calculate the cost of a NAV error.
For the error log: one row per error per account, the loss including costs, the policy rule applied, the date paid and the interest. In basis points these amounts are tiny, 1.09 bp of Fund A's NAV here. The finding, if there is one, will be about the rule, not the size.
Compute every error in every account, reimburse every loss with its costs and its delay, and treat netting as a disclosed exception rather than a default. The book's approach to turning fee and expense judgement calls into basis points is in the free workbook for this book.
Only where the written policy allows it and the fund documents disclose it, and many advisers do not allow it. Netting across different clients is never appropriate, because it uses one client's money to pay for another's loss. In the illustrative case netting across errors within Fund A cuts the adviser's payment from $48,650 to $38,650.
Under a no-netting policy the client account keeps it, because the account bore the risk of the error trade. In the illustrative quarter Fund A keeps 10,000 of gains from two errors while the adviser reimburses all 48,650 of losses across both funds.
Yes. The fund would not have paid them without the error, so they are part of the loss. In the illustrative fat-finger error, 45,000 excess shares cost 33,750 in price movement and 900 in commissions on the two legs, a reimbursable loss of 34,650.
This article is one calculation from The Private Fund Compliance Officer. 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: regulation, compliance and banking → · All 1501 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.