Four subsidiaries, twelve intercompany payments a month, and the arithmetic that shows which part of the saving is real.
Multilateral netting saves the cost of every intercompany payment that no longer has to be made: the FX spread, the transfer fee and the value days lost on the amount that offsets. On an illustrative group of four subsidiaries paying each other 54.0 million a month, netting reduces settlement to 6.5 million, an 88.0 per cent reduction, and saves 920,346 a year, or 770,346 after a 150,000 system cost. Of that saving, 92.9 per cent is the FX spread.
Worked in full in Treasury Management by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Four subsidiaries, each in its own currency, invoice each other for goods, services, royalties and management fees. All figures are in EUR millions a month and illustrative. Each row is what that entity pays to the others.
| Payer | To UK | To Germany | To US | To Poland | Total paid |
|---|---|---|---|---|---|
| UK | 8.0 | 5.0 | 2.0 | 15.0 | |
| Germany | 6.0 | 7.0 | 4.0 | 17.0 | |
| US | 3.0 | 9.0 | 1.0 | 13.0 | |
| Poland | 1.5 | 5.5 | 2.0 | 9.0 | |
| Total received | 10.5 | 22.5 | 14.0 | 7.0 | 54.0 |
The cost assumptions, also illustrative: a bank FX spread of 0.15 per cent on each cross-currency payment, a fee of 30 per cross-border transfer, one value day lost per payment at a 4.0 per cent funding rate, and 150,000 a year for the netting platform and the time to run it.
1. Each entity's net position. Receipts less payments, from the column and row totals.
Net position = total received − total paid
UK 10.5 − 15.0 = −4.5. Germany 22.5 − 17.0 = +5.5. US 14.0 − 13.0 = +1.0. Poland 7.0 − 9.0 = −2.0.
Excel, with payers in rows B2:E5: =SUM(B$2:B$5)-SUM(B2:E2) for each entity, using the matching column and row.
The positions sum to zero, which is the first check on any netting run. On netting day the UK pays 4.5 and Poland 2.0 into the netting centre, which pays Germany 5.5 and the US 1.0. Settlement is the sum of the positive positions: 6.5 a month instead of 54.0, in four payments instead of twelve.
2. Bilateral netting, for comparison. Offsetting each pair separately is what many groups already do informally. UK and Germany owe each other 8.0 and 6.0, so 2.0 moves. Across the six pairs the residuals are 2.0, 2.0, 0.5, 2.0, 1.5 and 1.0: 9.0 a month in six payments, an 83.3 per cent reduction.
3. Price each settlement method over a year. Twelve months of volume, the spread on every euro converted, a fee per payment and one day of funding on the amount in transit.
Annual cost = volume × spread + payments × fee + volume × days ÷ 365 × funding rate
Gross: 648,000,000 × 0.15% = 972,000; 144 × 30 = 4,320; float 71,014. Total 1,047,334.
Multilateral: 78,000,000 × 0.15% = 117,000; 48 × 30 = 1,440; float 8,548. Total 126,988.
| Method | Volume settled | Payments | FX spread | Fees | Float | Total |
|---|---|---|---|---|---|---|
| Gross, invoice by invoice | 648,000,000 | 144 | 972,000 | 4,320 | 71,014 | 1,047,334 |
| Bilateral netting | 108,000,000 | 72 | 162,000 | 2,160 | 11,836 | 175,996 |
| Multilateral netting | 78,000,000 | 48 | 117,000 | 1,440 | 8,548 | 126,988 |
Multilateral netting saves 920,346 a year against gross settlement: 855,000 of FX spread, 2,880 of fees and 62,466 of float. After the 150,000 system cost the group keeps 770,346, which is 0.119 per cent of the gross volume. Two things in that table are worth reading twice. The transfer fees, which netting projects are usually sold on, are almost nothing. And bilateral netting already captures 871,338 of the saving: the multilateral step adds 49,008 a year and 24 fewer payments. Its real advantages are elsewhere, in a single netting calendar, one FX desk dealing the net currency positions at a better spread than each subsidiary gets on its own, and a clean record of intercompany balances for the auditors.
| FX spread | Gross FX cost | Netted FX cost | Net saving |
|---|---|---|---|
| 0.05% | 324,000 | 39,000 | 200,346 |
| 0.10% | 648,000 | 78,000 | 485,346 |
| 0.15% | 972,000 | 117,000 | 770,346 |
| 0.25% | 1,620,000 | 195,000 | 1,340,346 |
The break-even spread is 1.5 basis points, so the spread rarely decides whether netting pays. Directionality does. Here net settlement is 12.0 per cent of gross because every entity both buys from and sells to the others. If it were 25 per cent, the saving at a 0.15 per cent spread would fall to 635,140; at 50 per cent, 374,387; at 75 per cent, 113,633. A group whose subsidiaries all sell into one distribution company has little to offset, and should look at cash pooling before netting.
Quoting the spread on the gross flows as the saving. On this group that gives 972,000, which forgets that the net positions still have to be converted, at 117,000, and that the system has a running cost. The true figure is 770,346, so the business case is overstated by 201,654, about a quarter. The second mistake is netting payments that should not be netted: flows into a country with exchange controls, or into an entity whose local rules require gross settlement, must be carved out before the netting ratio is measured, not after the project is approved.
Netting moves cash between entities, so it also changes where the cash sits on the morning of the netting date. That matters for the question the book is built around, how much of a group's cash is actually available, and it is why the netting calendar belongs in the 13-week forecast.
The book works through one company's five cash balances and the liquidity behind each, and the free workbooks for this case reproduce every figure it prints with a live formula.
Bilateral netting offsets each pair of entities separately; multilateral netting offsets every entity against all the others through a netting centre, so each settles one amount. In the illustrative group, bilateral netting cuts 54.0m of monthly flows to 9.0m in six payments, multilateral to 6.5m in four. The extra step is worth 49,008 a year here: most of the saving comes from the pairs.
Divide the system cost, less the fee and float savings, by the volume that no longer converts. With 570m of annual volume removed and 65,346 saved on fees and float, a 150,000 system breaks even at a spread of 1.5 basis points. Almost any bank spread on small intercompany tickets is above that, so volume and directionality decide it, not the spread.
Less and less. If net settlement is 25 per cent of gross flows, the saving after a 150,000 system cost falls to 635,140 on the illustrative group; at 50 per cent it is 374,387 and at 75 per cent 113,633. A group whose subsidiaries mostly sell to one distributor has little to offset.
This article is one calculation from Treasury Management. 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: corporate finance, valuation and markets → · All 1340 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.