BUILD: COURSE 3 | LESSON 3
Swaps and carry as strategy inputs
Learning objectives
Compute the swap cost or credit of any position from contract specifications, including triple-swap day and currency conversion
Integrate carry into strategy design: as a holding-cost budget for swing trades and as a directional filter or tailwind
Judge when swap costs disqualify a trade and how account type (including swap-free) changes the calculation
From footnote to line item
At Foundations level, swap was a footnote: "a small overnight fee". At Professional level it is a strategy input with a sign, a size, and a compounding schedule. The economics come from A1.1: holding a currency position overnight implies borrowing one currency to hold another, and the interest differential — adjusted by your broker's markup — is settled into your account daily as the swap (rollover). Long the higher-yielding side of a pair and you may earn; short it and you pay; and because the broker's markup is subtracted from both directions, plenty of pairs charge swap both ways.
The operational facts. Swap posts once per trading day at rollover (5pm New York). MetaTrader publishes per-instrument swap long and swap short values in the contract specification, usually in points per lot per night (sometimes in percent or currency — check the "swap type" field). And because spot FX settles T+2, positions held over the day that rolls across the weekend settlement — Wednesday for most FX pairs — are charged triple swap to cover Saturday and Sunday. (Many non-FX CFDs apply the triple on Friday instead; again, the specification is the source of truth.)
The arithmetic, end to end
Worked example. You are long 1 standard lot AUD/JPY. The contract specification shows swap long = +0.45 points, swap short = −9.8 points, swap type: points, where 1 point = 0.001 JPY (the 3rd decimal on a JPY pair quoted to 3 places).
- Nightly credit (long): 0.45 points × 100,000 units × 0.001 = 45 JPY ≈ $0.30 at USD/JPY 150. On Wednesday: ×3 = 135 JPY.
- Nightly cost (short): 9.8 points → 980 JPY ≈ $6.53 per night; ≈ $58.77 per week (7 nightly charges including the Wednesday triple: 9 charge-units × $6.53 — 9 nights' worth per 7-day week).
Note the asymmetry: the long earns a pittance while the short pays heavily — the markup ate most of the theoretical credit. That asymmetry is the broker's spread on interest, and it varies widely by pair and broker; never assume the credit side of a differential survives the markup without checking.
Now scale it against a trade plan. Suppose you short AUD/JPY as a three-week swing: 1 lot, 100-pip stop, and a JPY-pair pip worth ~$6.67 at USD/JPY 150 — risk ≈ $667. Swap over ~21 days (including three Wednesday triples) ≈ 27 charge-units × $6.53 ≈ $176 — roughly 0.26R of pure holding cost. Your trade must now clear a bar 0.26R higher than the chart suggested: a planned 2R winner is really 1.74R, and a breakeven scratch is a −0.26R loss. Over a year of such swings, carry drag of this size is frequently the difference between a profitable and losing book — invisible to anyone who only ever looks at charts.
The general formula to internalise:
with charge-nights = calendar nights + 2 × (number of triple-swap days in the window).
Carry as filter, tailwind and strategy
Once quantified, carry graduates from cost to input, in three escalating roles:
1. A veto/budget on swing trades. Add a line to every multi-day trade plan (A1.5's template): expected holding period × nightly swap, in R. Set a personal threshold — say, holding cost above 0.3R requires either a bigger expected move or a shorter plan. Direction matters: the same pair may be nearly free to hold long and ruinous to hold short.
2. A tiebreaker and tailwind. Between two comparable setups, prefer the one where carry pays you to wait. Positive-carry positions also forgive time: a range-bound week costs a negative-carry trade money and earns a positive-carry trade its keep. This shapes trade selection at the margin without ever being the reason for a trade.
3. The carry trade proper — with A1.1's warnings attached. Systematic carry harvesting (long high-yielders funded by low-yielders) is a real institutional strategy with a documented return profile: steady accrual punctured by violent, correlated unwinds — negatively skewed, crash-prone, and implicitly short volatility. A retail version layered on leverage magnifies both sides. If you pursue it at all: small size, wide stops or options-style predefined risk, regime awareness (volatility spikes and funding-currency central-bank turns are the eject signals), and the understanding that the swap credit is compensation for crash risk, not free money. This course does not present carry harvesting — or any strategy — as a system to copy; it presents the maths that lets you price what you are being paid to hold.
Account types and edge cases
Swap-free (Islamic) accounts replace overnight swap with either a flat administration fee after a grace period or wider costs elsewhere; for multi-week holds the comparison is a spreadsheet exercise, not a slogan — sometimes swap-free is cheaper for long negative-carry holds, sometimes markedly dearer. Swap values change — they track central-bank rates and broker funding conditions, so a position opened into a 2% differential can be sitting on a 0% differential after two policy meetings; re-check specifications monthly and around central-bank dates (your A1.5 calendar already flags them). Metals and index CFDs carry financing charges built the same way (funding rate ± markup); gold's swap is usually negative both ways, which is one reason multi-month CFD holds of gold underperform the metal itself. Finally, fold swap into your statistics: A2.3's expectancy should be computed on P&L net of swap, and A3.1's backtests must model it — a swing-strategy backtest without swap is systematically flattering the results.
Key takeaways
Swap = interest differential ± broker markup, charged per night at 5pm NY, tripled on the T+2 weekend-settlement day (Wednesday for most FX pairs)
Compute it from the contract specification: points × lot size × point value, converted to account currency — the credit side rarely survives the markup intact
Express holding cost in R (nightly swap × charge-nights ÷ risk): a 0.26R drag turns a 2R target into 1.74R and a scratch into a loss
Use carry as a veto/budget on swings, a tiebreaker between setups, and — only with full crash-risk awareness — a strategy component; never as free money
Swap values move with policy rates: re-check them monthly, model them in backtests, and compute expectancy net of swap