How margin differs across the asset classes

Three contracts of the same value 

The difference is clearest with everything held constant except the class. Three positions written on the same contract value, one on a broad index, one on gold, one on a single listed company. Contract sizes and prices differ, but by construction the notional values land in the same place, so the multiplication that produces the amount held has one remaining variable in it.

The second thing to hold constant is the market move, because the amount held is not what the profit and loss calculation reads. That calculation reads the contract value, which is the same in all three cases, so a move of a given percentage produces exactly the same money amount on all three positions.

Those last three rows are the whole of the class difference, read from the other end. A lower requirement percentage means a given amount held stands behind a contract value that a loss is calculated on in full, so a move of identical size consumes a much larger proportion of it, and the loss is not limited to the amount deposited; a favourable move of identical size is measured on exactly the same basis and produces the mirror figure. The percentage is not a fee and it measures nothing about the market. It states how much of a contract's value has to be present while the contract exists.

What the percentage is derived from 

A requirement percentage answers one narrow question the counterparty has to ask about every instrument it writes contracts on: if an account stops being able to support a position, how far can that position move between the moment this becomes true and the moment it is actually closed. The amount held is sized to cover that interval, and everything else about the instrument is irrelevant to the figure.

Two properties of the instrument determine how large that interval can be. The first is volatility, meaning how far the price of the underlying moves over a given period. A market that routinely travels a long way in an hour can travel a long way inside the seconds a close out takes, so more of the contract's value has to be sitting there in advance. Volatility is measured from observed price history, which is why requirements are usually revised after conditions change rather than in anticipation of them.

The second is liquidity, meaning whether a contract of the relevant size can be closed near the price on the screen. A closing instruction on a thin market moves the price it is filling against, so the exit is not at the level the calculation assumed. Liquidity is also the property that changes fastest: a market that is deep in ordinary conditions can thin out at the moment a great many positions need to be closed, which is when a close out is most likely to be running in the first place.

How each class behaves 

Major currency pairs sit at the low end on both inputs at once. They are the most heavily traded contracts in existence, quoted continuously through the trading week, and in percentage terms they move less over a day than most other markets. A currency contract is also the one case where the notional value does not move with the price, because the quantity is a fixed number of units of the base currency, so the amount held is unusually stable. That stability does not extend across the class. A pair involving a thinly traded currency can carry a requirement several times the one on a major, because both inputs move together: the book is thinner and the moves are larger.

Major indices sit close to the currency pairs, for a structural reason. An index is an average of many companies, so the idiosyncratic moves of its constituents partly cancel and the average travels less than the shares inside it. What an index contract does carry is a gap between sessions. The constituents stop trading when their exchange closes while news does not stop arriving, so the first price of the next session can be some distance from the last price of the previous one, and a gap is a move no close out can be executed inside.

Commodities and metals sit above the indices as a group and hold the widest spread of behaviour in the class list. Gold is quoted almost continuously into a deep book and behaves, on the liquidity input, more like a major currency pair than like the rest of its own class. An energy contract responds to inventory reports, production decisions and physical disruption, and its price history contains moves no equity index has ever produced. Agricultural contracts answer to weather and harvest calendars. A single requirement percentage for the class as a whole would be meaningless, and no venue publishes one.

Shares CFDs sit at the high end, because the averaging that holds an index down is absent. One company can move a very long way on one announcement, with nothing to absorb it: results, guidance changes, regulatory decisions and corporate actions all land on a single instrument. The trading day is short as well. A listing is open for a fraction of the day, so a shares position spends most of its life unable to be closed at all, and the requirement has to stand against whatever the price does before the next opening.

Exchange traded fund CFDs sit across the widest band of all, because the class name describes a wrapper rather than an exposure. A fund tracking a broad developed market holds hundreds of companies and behaves, on the volatility input, very much like an index. A fund tracking one sector, one country or one commodity carries the concentration of the thing it tracks. The fund's own market matters too: the wrapper is a listed security with its own liquidity regardless of how liquid its holdings are, and a small or specialised fund can be thin where a large established one is not.

The class is a hypothesis, not an answer 

The ordering above describes how class averages are usually arranged, and there is no such thing as an average instrument. The spread of requirements inside a single class routinely exceeds the gap between one class average and the next: a major pair and a thinly traded one, a large listed company and a small one, a broad market fund and a single sector fund. Two instruments filed under the same heading on the same platform can carry requirements that are multiples apart, and nothing about the heading predicts which.

The class is therefore useful for understanding why the numbers differ and useless as a source for any particular number. It explains the shape of the schedule. It does not supply the figure for a named instrument, which is published, specific, and frequently some distance from what the class average would suggest.

When a published requirement changes 

A published requirement is a current term rather than a permanent one, and the commonest reason it changes is a known event with a date on it. A scheduled central bank decision, a company results announcement, a national vote: all share one property. Everyone knows when the information arrives, nobody knows what it says, and the first price afterwards can be a long way from the last price before. Counterparties commonly raise the requirement on the affected instruments beforehand and restore it afterwards.

The effect on an account is worth working through, because it arrives without anything happening to the position at all.

Nothing was traded in that block. No price moved, the position was never touched, the equity never changed, and the margin level halved anyway. This is the one change to an account panel that originates entirely outside the account, and it is the reason a requirement calculated once at the moment a position opened does not stay authoritative on its own.

Weekends work the same way, and the reason is not that weekends are dangerous. It is that a closed market cannot be closed out of. Information keeps arriving while a venue is shut, and price discovery resumes at the reopening rather than picking up where it stopped, so a position carried across a closure is exposed to a distance no close out can be executed inside. Requirements are therefore commonly raised before the last session of the week and restored after the first session of the next. Long public holidays get the same treatment, and so does the run up to expiry on contracts written on futures, where the underlying contract is being rolled and the book around it thins.

Finding the figure for a named instrument 

Every instrument carries a contract specification, the published document of record for its terms. It is the only place the operative requirement exists. Reading one means locating a small number of fields, in this order, because each depends on the one before it.

1. Contract size, the quantity of the underlying that one contract covers: units of a base currency, ounces, a money amount per index point, or a number of shares or fund units.
2. The currency the contract is valued in, which determines whether a conversion step stands between the calculation and the figure the account panel reports.
3. The requirement itself, and critically its form: a percentage of the contract value, or a fixed money amount per contract that does not move when the price does.
4. Whether the requirement is tiered by position size, in which case the specification carries a table of size bands and the figure changes as a position crosses a band rather than staying flat.
5. Trading hours, plus any stated weekend, holiday or scheduled event treatment, since those are the conditions under which the published figure is temporarily replaced by a different one.
6. Expiry, on any contract written on a future, together with the venue's stated treatment as that date approaches.

A YAL account runs on one of two platforms, MetaTrader 5, and both publish these fields per instrument rather than in one schedule. On MetaTrader 5the specification window for a symbol lists the contract size, the valuation currency and the requirement alongside the rest of the terms. The order ticket then reports the amount that would be held for the size entered, which is the same arithmetic reading the same published figure.

A specification read once is a snapshot of a term the venue can revise, including on a position that is already open, so the figure governing an open position is the one currently published rather than the one that applied when it opened. That holds for every class in this lesson, and it is the single point on which they do not differ at all.

Where practitioners disagree 

The first disagreement is about event and weekend re-rating itself. One tradition holds that gap risk is entirely real, that a requirement ignoring it understates what the instrument costs to stand behind, and that raising the figure before a known event is honest arithmetic rather than a penalty. The other holds that revising a requirement on an already open position moves the consequences of the firm's own risk management onto the holder at the least convenient moment, and that a term the venue may change at its discretion is materially weaker than a fixed one. Both are in wide use, disclosure practice varies considerably between venues, and neither argument has retired the other.

The second concerns exchange traded funds, and it is part of why that class shows the widest dispersion. One view treats a broad fund as the basket it holds, on the grounds that the diversification inside it is genuine and its volatility is measurably closer to an index than to a share. The other treats it as a listed security in its own right, on the grounds that it has its own book, its own trading hours and its own tracking behaviour, so its price can move for reasons the underlying basket did not produce. The consequence is that two funds offering similar exposure at different venues can carry noticeably different requirements, and the gap is a difference of methodology rather than an error by either.