Where does leverage actually come from?
Leverage is usually explained as "borrowing to invest". In derivatives it arrives without anybody borrowing anything, and understanding why is what makes the risk legible instead of mysterious.
The mechanism, in one paragraph
You never buy the underlying. You agree to a payoff calculated on the underlying. Since nothing is purchased, no purchase price is required. But the clearing house still needs assurance you can pay tomorrow's losses, so it demands a deposit — initial margin — sized to cover a plausible adverse move over a short close-out window, not to cover the position's value. Cover a couple of days' plausible movement instead of the full notional and you have created leverage as a by-product of a sensible risk-management rule.
Put the numbers on a single contract
One crude oil futures contract is 1,000 barrels. At $80, its notional is $80,000. Suppose initial margin is $6,000 (illustrative; exchanges reset it as volatility changes).
Leverage L = notional ÷ capital = $80,000 ÷ $6,000 ≈ 13:1
Now move the price:
| Oil price | Move | Position P&L | Effect on the $6,000 |
|---|---|---|---|
| $84.00 | +5% | +$4,000 | +67% → $10,000 |
| $76.00 | −5% | −$4,000 | −67% → $2,000 |
| $72.00 | −10% | −$8,000 | −133% → −$2,000 |
Read the last row carefully. An 8,000 dollar loss against a 6,000 dollar deposit leaves the account owing $2,000. The position did not stop at zero. A leveraged derivative can lose more than the amount put in, and a 10% move in crude oil is an entirely ordinary event, not a tail scenario.
Some jurisdictions require brokers to give retail clients negative-balance protection, which caps the loss at the deposit. That is a consumer-protection rule in specific markets — not a property of derivatives, and not available to professional or institutional accounts.
The two formulas worth memorising
Equity change = L × underlying change. At 13:1, a 3% move is a 39% swing in your capital. At 30:1 — a common cap for retail currency products — a 3% move is a 90% swing.
Ruin point = 1 ÷ L. An adverse move of 1/L wipes out the capital entirely.
| Leverage | Adverse move that erases the deposit |
|---|---|
| 2:1 | 50% |
| 5:1 | 20% |
| 13:1 | 7.7% |
| 30:1 | 3.3% |
| 100:1 | 1.0% |
At 30:1, a 3.3% move ends the position. In a major currency pair a single session that large is rare — EUR/USD has not closed 3% away from the previous close once in the past decade, the 2022 rate cycle and the April 2025 tariff week included; the largest in that window was 2.3%. What is ordinary there is a 1% day, roughly one session in thirty, and at 30:1 that alone is a 30% swing in the deposit. The 3.3% day still arrives, without warning, when a central bank moves or a referendum lands; individual shares do it on earnings.
Margin calls make the timing worse
Below the initial margin sits a maintenance margin. Fall below it and cash must arrive, usually the same day. If it does not, the broker closes the position at whatever price exists at that moment — not the price you would have chosen, and typically in the middle of the move that caused the call. Leverage therefore does not merely multiply outcomes; it removes your ability to wait, which is the one thing an unlevered holder always retains.
What the record shows
Stated as documented observation rather than as a warning: EU and UK regulators require providers of retail CFDs and similar leveraged products to publish the share of retail accounts that lose money, and those disclosures have commonly sat in the 70s and 80s percent, with ESMA's own pre-2018 samples reporting roughly 74% to 89% of retail accounts losing money. Academic studies using complete brokerage records of retail day traders — the large Taiwanese and Brazilian samples are the best known — have found only a small minority consistently profitable net of costs. These describe populations, not individuals, and they predict nothing about any particular person. They are the reason this course teaches the arithmetic rather than the products.
Try it now
- A year of daily candles on a major currency pair is below. Measure a handful of sessions to get a feel for what an ordinary day looks like in percentage terms.
- Compute 1/L for L = 5, 13 and 30 — the move that wipes out the equity at each leverage. Then scan the year for days that exceeded each of the three. Three counts, three very different pictures of the same instrument, and the instrument did not change.
- Measure the single worst day in the window and compute the equity outcome at each of those three leverage levels, letting the result go negative where it does. Nothing here recommends any leverage level, including none — it shows what the multiplier does to a number you already had.