Pro rata to the pre-trade order, rounded to the increment, one average price: and the minimum-piece rule and fill sequence that quietly move money.
Allocate a partial fill in proportion to each account's pre-trade order, round each share to the instrument's trading increment, apply the policy for any account that falls below the minimum piece, and give every account the same average price. On an illustrative 10,000,000 bond order filled at 5,800,000, the two funds receive 4,186,000 and 1,614,000 at an average price of 98.564, and the 300,000 separate account receives nothing, because its pro rata share of 174,000 is below the 200,000 minimum. Handing the fills out in time order instead would have moved 6,884 of cost from one fund to the other.
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 →
Partial fills are routine for credit managers buying bonds and syndicated loans for several funds at once: the order goes in for the whole complex, the market delivers part of it, in pieces, at different prices. The allocation is a conflict the examiner will look at, because the adviser chooses who gets the cheap pieces. The standard most firms follow comes from the SEC staff's 1995 SMC Capital no-action letter on aggregated orders: a written allocation statement before the trade, a partial fill allocated pro rata to that statement, and every participating account receiving the average price. The arithmetic below shows where the judgement calls hide.
| Input | Value |
|---|---|
| Fund I order (face) | 7,000,000 |
| Fund II order | 2,700,000 |
| Separately managed account (SMA) order | 300,000 |
| Block order | 10,000,000 |
| Fill 1 | 2,400,000 at 98.25 |
| Fill 2 | 2,000,000 at 98.60 |
| Fill 3 | 1,400,000 at 99.05 |
| Total filled | 5,800,000 |
| Minimum piece / trading increment | 200,000 / 1,000 |
The orders were recorded on the allocation statement before the first fill. That timestamp matters more than anything that follows: an allocation decided after the prices are known is not an allocation policy, it is a choice.
Step 1, fill ratio. 5,800,000 ÷ 10,000,000 = 58.0 per cent.
Step 2, raw pro rata shares. Fund I 7,000,000 × 58.0% = 4,060,000. Fund II 2,700,000 × 58.0% = 1,566,000. SMA 300,000 × 58.0% = 174,000.
Step 3, minimum piece. 174,000 is below the 200,000 minimum the bond can be held in, 26,000 short. The SMA cannot receive its share.
Step 4, reallocate among eligible accounts. The policy says a share below the minimum piece is zeroed and the fill is reallocated pro rata among the others: Fund I 7,000,000 ÷ 9,700,000 × 5,800,000 = 4,185,567; Fund II 1,614,433.
Step 5, round to the increment. Round down to 1,000 and give the residual to the largest remainder: Fund I 4,186,000, Fund II 1,614,000, total 5,800,000. Both funds end at 59.8 per cent of their order.
Step 6, average price. (2,400,000 × 98.25 + 2,000,000 × 98.60 + 1,400,000 × 99.05) ÷ 5,800,000 = 98.5638, consideration 5,716,700 before accrued interest.
In Excel, with orders in B2:B3 and the total fill in B6: =FLOOR(B2/SUM($B$2:$B$3)*$B$6, 1000), then the residual increments by remainder; average price =SUMPRODUCT(Q,P)/SUM(Q).
| Account | Order | Raw pro rata | Allocated | Fill % | Cost at 98.564 |
|---|---|---|---|---|---|
| Fund I | 7,000,000 | 4,060,000 | 4,186,000 | 59.8 | 4,125,880 |
| Fund II | 2,700,000 | 1,566,000 | 1,614,000 | 59.8 | 1,590,820 |
| SMA | 300,000 | 174,000 | 0 | 0.0 | 0 |
| Total | 10,000,000 | 5,800,000 | 5,800,000 | 58.0 | 5,716,700 |
The compliance test on this ticket is short: each eligible account's allocation is within one increment of its pro rata target (Fund I is 433 above, Fund II 433 below), the price is identical across accounts, and the zeroed SMA is explained by a written rule rather than by a trader's note.
Some policies round a small account up to the minimum piece instead of zeroing it. Then the SMA takes 200,000 and the remaining 5,600,000 is split pro rata: Fund I 4,041,000, Fund II 1,559,000. The SMA ends at 66.7 per cent of its order while the funds get 57.7 per cent. Both rules are defensible; what is not defensible is choosing between them ticket by ticket.
| Fill ratio | Filled | Fund I | Fund II | SMA | SMA at minimum? |
|---|---|---|---|---|---|
| 40% | 4,000,000 | 2,800,000 | 1,080,000 | 120,000 | No |
| 58% | 5,800,000 | 4,060,000 | 1,566,000 | 174,000 | No |
| 70% | 7,000,000 | 4,900,000 | 1,890,000 | 210,000 | Yes |
| 90% | 9,000,000 | 6,300,000 | 2,430,000 | 270,000 | Yes |
The SMA participates only when the fill reaches 66.7 per cent. A small account that is zeroed on every scarce new issue and filled on every easy one is being systematically disadvantaged by a neutral rule, and the annual review should count how often it happens.
The more expensive shortcut is allocating fills in the order they arrive. If Fund I takes the first fills until it reaches 4,186,000, and Fund II gets what is left, the quantities are still pro rata but the prices are not:
| Sequence | Account | Price paid | Gain (−) or cost against average | Basis points |
|---|---|---|---|---|
| Fund I first | Fund I | 98.399 | −6,884 | −16.7 |
| Fund I first | Fund II | 98.990 | 6,884 | 43.3 |
| Fund II first | Fund II | 98.250 | −5,065 | −31.8 |
| Fund II first | Fund I | 98.685 | 5,065 | 12.3 |
With 0.80 points between the first and last fill, the account served second overpays by 12 to 43 basis points on a single ticket. Repeated across a year of rising markets, the account that is habitually first in the queue, often the one paying higher fees or carrying the firm's own money, collects a transfer that no single trade makes visible.
Testing quantities and not prices. Most allocation reviews compare each account's fill percentage with the target and stop. That test passes the time-order allocation above: both funds received exactly 59.8 per cent. Only a price test, the account's average cost against the block's average price, catches it. The second mistake is the order record itself. If the allocation statement is written up after the last fill, or amended after execution without a documented reason, the pro rata arithmetic is applied to a target that was itself chosen with hindsight. Sample the timestamps as well as the numbers; how many items you need to test works the sample size, and the same reasoning applies to deal-level allocation in how to test whether deal allocations favour one fund.
The testing logic and the fee and expense workbook behind this programme are in the free workbook for this case.
It cannot be allocated as computed, so the policy must say in advance whether the share is zeroed and reallocated to the other accounts, or rounded up to the minimum and taken from them. In the illustrative case the SMA's 174,000 share is 26,000 below the 200,000 minimum: zeroing it gives the funds 59.8 per cent of their orders, rounding up gives the SMA 66.7 per cent and the funds 57.7 per cent.
That is the standard most advisers adopt from the SEC staff's 1995 SMC Capital letter: each participating account receives the average price of the block. In the illustrative case three fills at 98.25, 98.60 and 99.05 average 98.564. Allocating the fills in time order instead would have given one fund 98.399 and the other 98.990, a transfer of 6,884 on a single ticket.
Test both quantity and price. Check each account's allocation is within one trading increment of its pro rata target, in the example 433 either way on a 1,000 increment, then check each account's average cost against the block average. A quantity-only test passes a time-order allocation where both funds got exactly 59.8 per cent but paid prices 0.59 points apart.
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 1682 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.