Ask a retail investor what leverage their broker offers and you will get a single figure: 1:30, 1:100, 1:500. Ask what leverage they were actually given on their last large position during a volatile session, and most have no idea — because the answer is often not the number on the account page.
Over the past decade the industry has quietly moved from static leverage, set once per account, to tiered and rule-driven margin, where the ratio applied to a position depends on its size, on account equity, on the instrument, and on what the calendar says is about to happen. The dynamic leverage systems that brokers use to enforce this operate silently in the background, recalculating margin requirements as positions grow. For an investor, the practical consequence is that the leverage figure in the marketing material is a ceiling, not a constant — and understanding the difference is the difference between a controlled drawdown and a stop-out you did not see coming.
Three things that get confused with each other
Precision here saves a lot of trouble later.
Leverage is a ratio describing how much notional exposure you may control per unit of your own capital. At 1:100, $1,000 of margin supports $100,000 of notional exposure.
Margin is the capital actually locked against an open position. It is the operational quantity — the one your platform checks, the one that triggers a margin call. Margin required equals notional divided by leverage.
Risk is how much you lose if the market moves against you. Leverage does not create risk directly. Position size creates risk. Leverage determines how large a position your capital permits, and therefore how large a position you are tempted to take.
This last distinction is the one most commonly missed. A trader holding one mini lot of EURUSD faces exactly the same profit and loss per pip whether the account is set to 1:30 or 1:500. What changes is the free margin remaining, and therefore how much adverse movement the account can absorb before the broker begins closing positions. High leverage does not make a trade riskier. It makes a larger trade possible, and it reduces the cushion protecting the position you have.
Why one static ratio was always a blunt instrument
Fixed leverage treats a $500 account and a $500,000 account, a quiet Tuesday and a central bank decision, a micro position and a fifty-lot position, as the same problem. They are not.
Consider the arithmetic. A $10,000 account at 1:500 can open roughly $5,000,000 of notional exposure — about fifty standard lots of a major currency pair. At that size, a ten-pip adverse move is a $5,000 loss: half the account. A twenty-pip move ends it. Ten and twenty pips are ordinary noise on any major pair; they happen several times a day without anything notable occurring.
So at the extreme end, high static leverage does not offer meaningful opportunity. It offers a route to a stop-out within minutes of ordinary market movement. This is the reason the leverage question became a regulatory issue at all, and the reason several major jurisdictions capped retail leverage rather than leaving it to firms to decide.
But a blanket cap is also blunt. It treats a highly liquid major pair the same as a thinly traded emerging-market cross whose spread can multiply overnight. It applies the same limit to a trader holding a hedged pair of positions for two minutes and to one holding directional exposure across a weekend. And it does nothing at all about the specific moments — scheduled data releases, elections, central bank meetings — when the probability of a violent gap rises by an order of magnitude.
Tiered and dynamic margin exists because both the single account-level ratio and the single regulatory cap are too coarse to manage risk that varies along several dimensions at once.
How tiered margin actually works
The underlying principle is straightforward: leverage decreases as exposure increases.
A typical tier table for a major currency pair might look roughly like this. The first few standard lots of notional exposure receive the account’s headline leverage. The next band receives half of it. Beyond that, a quarter. At the top band — genuinely large exposure for a retail account — leverage may fall to 1:10 or lower.
Two implementation details determine how this feels in practice.
Marginal versus blanket application. Under marginal application, each band is margined at its own rate, like a progressive tax: your first tranche keeps the high ratio even after you enter the next band. Under blanket application, crossing a threshold re-margins your entire position at the lower ratio. The second is far more aggressive: adding one lot at the wrong moment can multiply the margin requirement on everything you already hold, and can trigger a margin call by itself. Brokers do not always make clear which method they use. It is worth asking.
Aggregation scope. Are tiers calculated per position, per instrument, per correlated group, or across the whole account? A trader holding several positions in correlated pairs may be carrying a single concentrated exposure while the platform, if it tiers per instrument, sees several small independent ones. More sophisticated risk engines aggregate by currency exposure rather than by symbol, which is materially safer for the trader as well as the broker.
Equity, not balance
A detail that sounds technical and turns out to be decisive: does the margin engine read account balance or account equity?
Balance is the settled cash in the account. Equity is balance plus or minus the floating profit and loss of open positions. A trader with $10,000 balance and $4,000 of unrealised losses has $6,000 of equity.
If the leverage tier is determined by balance, that trader still enjoys the leverage appropriate to a $10,000 account while actually controlling $6,000. If it is determined by equity, leverage steps down as the account deteriorates — which is precisely when a smaller exposure is appropriate.
Equity-based calculation is the more conservative and, for the investor, the more protective design. It has an uncomfortable property worth understanding: in a sharp adverse move, falling equity can push the account into a lower leverage tier, which raises the margin requirement, which reduces free margin further, which accelerates the approach to stop-out. The mechanism is procyclical by construction.
That is not a flaw so much as a trade-off made explicit. The alternative — holding leverage constant while equity collapses — ends in the same place, later, and frequently past zero. But it does mean that an investor holding a leveraged position into a deteriorating market should expect the rules to tighten as the position worsens, not stay still. Planning for a stop-out at the level implied by the opening margin requirement will produce an unpleasant surprise.
Scheduled risk: the calendar as a risk parameter
The most useful feature of rule-driven margin, and the one investors notice most, is scheduling.
Certain moments carry structurally elevated gap risk: non-farm payrolls, CPI releases, central bank rate decisions, national elections and referendums, and the Friday-to-Sunday weekend gap. Liquidity thins, spreads widen, and the probability of price jumping past the stop-loss level rather than trading through it rises sharply.
Brokers respond by tightening margin ahead of these windows — commonly a few hours before a major release, and from Friday afternoon into the weekend close. Positions opened before the change are usually margined under the new rules too, which means free margin can fall without the trader doing anything at all.
This is the single most common source of surprise margin calls among retail investors, and it is entirely avoidable. Brokers publish these changes in advance, typically in a notices section or by email. Very few clients read them. If you routinely hold leveraged positions over weekends or through scheduled data, finding the page where your broker announces margin changes and checking it weekly is perhaps the highest-value five minutes in retail risk management.
Why brokers on both sides of the book want this
It is tempting to read tightening margin rules as brokers protecting themselves at clients’ expense. The incentive structure is more interesting than that.
A B-book broker holds client risk on its own balance sheet. When leveraged clients blow up in a violent move, two things happen: the broker absorbs the immediate position risk, and any account that goes past zero becomes an uncollectible debit the broker must absorb under negative balance protection rules. The broker’s exposure is therefore concentrated precisely in the accounts running the most leverage into the most volatile moments. Reducing those clients’ leverage automatically is direct self-protection.
An A-book broker hedges externally and does not profit from client losses. Its revenue is a function of volume over time, which means its worst outcome is a client who deposits, over-leverages, is stopped out in a week and never returns. Client survival is the business model. Automatic leverage reduction extends account lifespan, and extended lifespan is exactly what an A-book broker is optimising for.
There is also a flow-quality dimension. Some traders deliberately hold offsetting exposure across several brokers, so that each individual broker sees a one-sided position while the trader carries little net market risk. To the broker, that flow looks like an ordinary directional client and is priced accordingly, until it isn’t. Exposure-sensitive margin rules limit the damage that kind of flow can do without requiring the broker to make case-by-case judgements about individual clients.
The commonality across all three cases is that mechanical, pre-announced, uniformly applied rules are better for clients than discretionary intervention. A broker that reduces leverage automatically according to a published table is behaving predictably. A broker that calls a profitable client to renegotiate their terms after the fact is not. Infrastructure vendors in this space — firms such as Takeprofit Tech that build the risk and execution layer brokers run on — essentially sell the ability to make risk management automatic rather than discretionary, which is why the arrival of these tools across the industry has been, on balance, good for retail clients even though it usually reduces the leverage available to them.
Negative balance protection and where the losses actually land
In several major jurisdictions retail clients cannot lose more than their deposited capital. If a gap takes an account below zero, the broker absorbs the shortfall.
This is a genuine and valuable protection, and it is also the reason leverage caps and dynamic margin exist in the same regulatory frameworks. The three are a package: if the broker must eat negative balances, the broker must be allowed — in fact required — to constrain the leverage that produces them.
Two practical notes. First, protection is a function of the entity you contract with, not the brand. Large groups operate multiple entities across jurisdictions, and clients are sometimes onboarded to an offshore entity offering higher leverage without negative balance protection, under the same logo and website. The entity name is in the client agreement. Read it.
Second, protection applies to the account, not to the position. It prevents you owing money you never deposited. It does nothing to prevent the deposit itself from being lost.
Turning all of this into position sizing
For an investor, the useful output of understanding margin mechanics is a sizing discipline that does not depend on the leverage available.
Size from risk, never from available margin. Decide the maximum you are prepared to lose on a trade — a fixed percentage of equity, commonly in the range of half a percent to two percent. Divide that by the distance in points to your stop level, and the result is your position size. Leverage enters this calculation only as a constraint on whether the resulting size is permitted at all. If your sizing is correct, the leverage on the account is usually irrelevant.
Model the gap, not the stop. Stops are not guarantees. In the moments dynamic margin is designed for, price can jump straight past your level. A guaranteed stop, where offered, costs a premium and converts an unknown into a known. For positions held across weekends or through major scheduled events, that premium is often the cheapest insurance available.
Keep free margin well above the margin call level. Treat the margin call level as a catastrophe threshold, not an operating range. A useful habit is to check what your margin level would be after an adverse move of, say, three times the average daily range — and to size so that even that scenario leaves the position intact.
Aggregate your correlated exposure manually. If your platform tiers margin per symbol, do the aggregation yourself. Several positions in pairs sharing a common currency are one position wearing several hats.
Check the tier table before scaling up. If your broker uses blanket rather than marginal tiering, know where the thresholds sit before you add to a winner. Crossing one accidentally is an unforced error.
One further habit is worth adopting: rehearse the arithmetic before the position exists. Take the size you intend to open, look up the tier that size falls into, calculate the margin it will consume, subtract it from your equity, and then ask what adverse move would take the remaining free margin to zero. If that number is smaller than a normal day’s range for the instrument, the position is too large regardless of what the platform allows you to open. The platform enforces the broker’s risk limits, not yours, and the two are set with entirely different objectives in mind.
Reading the rules before you need them
Everything discussed here is disclosed somewhere. It is rarely disclosed prominently. Before committing capital, locate and read:
- The margin and leverage schedule, including tier tables per instrument class and whether tiering is marginal or blanket.
- The statement of whether tiers are computed on balance or equity, and whether aggregation is per position, per symbol or per account.
- The notices page where scheduled margin changes are announced, and the notice period given.
- The margin call and stop-out levels as percentages, and the order in which positions are closed at stop-out — largest loss first, largest margin first, and other conventions produce materially different outcomes.
- The contracting entity and whether negative balance protection applies to it.
If any of these cannot be found in writing, that is itself an answer. A firm that has built a defensible rule set is generally willing to publish it; a firm relying on discretion tends to prefer vagueness.
Where the rules differ by instrument
Leverage is not a single policy applied uniformly across a broker’s product range. It is usually a matrix, and the differences between cells are informative about where the firm believes its risk sits.
Major currency pairs carry the highest available leverage almost everywhere, because they are the deepest and most continuously traded markets in existence. Minor and exotic pairs step down, sometimes sharply — an emerging-market cross can see its spread multiply during a local political event while the majors barely move.
Index CFDs sit somewhere in between, with the added complication that the underlying futures market has its own trading hours. A position held through a period when the underlying is closed carries gap risk that no stop can manage, which is why index margin frequently tightens outside core hours.
Commodities split by behaviour rather than category. Gold is treated almost as a currency by many brokers. Energy contracts are treated with far more caution, for the straightforward reason that the market has demonstrated within recent memory that prices can move by percentages that most risk models did not contemplate.
Crypto CFDs are the strictest category almost universally, and reasonably so. A market that trades continuously, has no circuit breakers, no central authority and a documented history of double-digit hourly moves is not a market where high leverage is survivable. Leverage in the region of 1:2 to 1:5 is common for retail crypto exposure in regulated jurisdictions, and offers advertising far higher figures are generally coming from entities where negative balance protection does not apply.
The practical takeaway is to check the specific instrument rather than the account headline. A trader who has internalised “my account is 1:200” and applies that assumption to a crypto or exotic position will discover the real number at the least convenient moment.
A worked example
A trader holds $20,000 of equity and a headline leverage of 1:200 on EURUSD, under a marginal tier structure: 1:200 up to $500,000 of notional, 1:100 from $500,000 to $2,000,000, 1:50 above that. Margin call at 100 percent, stop-out at 50 percent.
Opening ten standard lots — roughly $1,080,000 of notional at a rate of 1.08 — the first $500,000 requires $2,500 of margin, and the remaining $580,000 at 1:100 requires $5,800. Total margin used: $8,300. Free margin: $11,700. The margin level sits comfortably above 200 percent.
The trade moves against the trader by 80 pips. Floating loss is $8,000, equity falls to $12,000, and the margin level drops to roughly 145 percent. Still open, but the cushion has more than halved on a move that is well inside a normal day’s range.
Now a scheduled event arrives and the broker halves leverage across the board for the duration. Margin required doubles to $16,600 against equity of $12,000. The margin level falls below 75 percent, the margin call has already passed, and the account is one small further move from stop-out — without the trader having placed a single additional order.
Nothing irregular occurred here. The tier table was published, the event change was announced in advance, and every calculation followed disclosed rules. The position was simply sized as though the opening margin requirement was the operative constraint, when in fact it was the loosest constraint the position would ever face.
That is the whole lesson in one example. Leverage is no longer a number attached to an account. It is a function — of size, of equity, of instrument and of time — and the only reliable way to trade around it is to size positions so that the function’s output never becomes the binding constraint on whether your trade survives.