Margin and Leverage

Margin is the collateral you post to hold a position. It is not a measure of what the position can lose — the two numbers are set by different people for different reasons — and sizing to the first instead of the second is the most common way an account ends. On the contract worked below, the deposit the broker asks for is fifty times the money the trade actually risks.

Prerequisites: Lesson 13, for which of the four objects have margin at all, and lesson 10, for the four charges that margin is not one of.

One contract, five deposits

Take the S&P 500 at 5,000 and the Micro E-mini, which is worth $5 a point. One contract therefore controls 5,000 × $5 = $25,000 of index. That $25,000 is the notional, and it does not change with your deposit. Only the leverage does.

Equity behind it	Leverage	Move that takes it all	S&P points
$25,000	1×	100%	5,000
$5,000	5×	20%	1,000
$2,500	10×	10%	500
$1,250	20×	5%	250
$500	50×	2%	100

Read the last column rather than the first, because it is in the units the market moves in. At fifty times leverage the deposit is gone on a hundred S&P points, and a hundred points is two average days of the range you measured in lesson 12. The instrument has not become more dangerous anywhere in that table; the same contract appears on every row. What changed is how much of the market’s ordinary movement you can absorb before you are out of money.

Now put the other number beside it. Suppose you trade that contract with a ten-point stop. Ten points at $5 is $50, so the trade risks $50 against a deposit of $2,500 — two per cent of the deposit and two tenths of one per cent of the notional. The margin is fifty times the risk. Neither number tells you the other, and only one of them is under your control.

That last sentence is the whole lesson. The margin was handed to you; the stop is yours. A trader who sizes to the margin has let the clearing house choose the position, and the clearing house was not trying to keep them solvent — it was trying to keep itself whole.

What this does not settle

That leverage is bad. Leverage is a ratio and ratios do not have intentions; the same 10× is reckless on a position you cannot watch and unremarkable on a hedged one held for an hour. What the lesson establishes is that the ratio is computable in advance, which is the only thing that lets you argue about it honestly.

Nor is the wipeout move in the table a forecast. It assumes you do nothing at all — no stop, no exit, no deposit — and so it marks a boundary rather than a probability. What actually happens as you approach that boundary is lesson 15, and how likely you are to reach it is lesson 22.

And margin does not bound your loss from below either. Because the requirement is collateral rather than a limit, a market that gaps through your stop can leave you owing more than you posted. The bond protects the clearing house against you, not you against the market.

Every rule quoted here is also a United States rule. Regulation T, the maintenance level and the day-trade courtesy are the regime of one jurisdiction, and elsewhere the initial requirement, the level at which someone else takes over, and whether you can be left owing money at all are all decided differently. Several jurisdictions give retail accounts statutory protection against a negative balance, which removes the last of the three bounds above entirely. The arithmetic in this lesson travels. The thresholds do not.

And the table is a photograph taken at the entry. Leverage is notional over equity and both move, so it climbs as a position goes against you: the ten-times row, after a five per cent fall, is running nineteen times, because the equity halved while the notional barely moved. The wipeout column is still exactly right, since the arithmetic is linear in points. What the table cannot show is the leverage you are actually carrying at the moment somebody calls you, which is well before the last row and is not a number this page contains.

The number your platform shows you before a trade and the number that decides whether you survive it are different numbers. Write both down.

Problems

1. Your actual leverage, five times over. For your last five positions, compute the notional value — contracts times point value times price, or shares times price — and divide by the equity in the account at the time. You now have five numbers you have probably never seen. Note the largest.
2. The move that would have taken the account. For that largest number, compute 1 ÷ L and turn it into the instrument’s own units. Then take the daily range you measured in lesson 12 and ask how many ordinary days that move is. If the answer is less than one, the position could not have survived a Tuesday, and you were relying on it not being one.
3. Put the two numbers side by side. For one trade you are actually planning, write the margin requirement and the stop-loss in dollars, then divide the first by the second. Do it again with the position doubled. One of the two numbers changes what you can lose and the other does not, and the exercise is worth nothing until you can say which without checking.

Sources. CME Group, performance bond requirements and the SPAN margin methodology, which sets the requirement from the instrument’s modelled volatility rather than from any participant’s intentions. Board of Governors of the Federal Reserve System, Regulation T, the initial margin rule for securities accounts, and FINRA Rule 4210, which sets the maintenance level below which a position is no longer yours to manage. The first governs futures and the other two govern securities, which is why the same money buys such different positions in the two. All three are worth reading once for the same reason: not one of them mentions your stop.

You can now compute, before opening a position, the move that would end it. The next lesson takes that boundary and puts a price on it, because someone else is watching it too, and they act first.

Educational only. Trading involves substantial risk of loss. Not financial advice. Past performance does not guarantee future results.