Position Limits: Caps on How Large a Position May Be

A position limit is a ceiling on how much of a single contract one participant may hold. Collateral asks whether a promise can be collected at all. A position limit asks something collateral never reaches: how large a promise this arrangement is prepared to have to collect from any one party. Where each ceiling sits, and what it is measured against, is set by the authority.

Collateral makes each rupee of a promise collectable. Collateral does not notice how many rupees there are, and was never built to. When a holding doubles, the collateral behind it doubles too, and every ratio computable from the pair comes out exactly where it was before. A holding can therefore be perfectly collateralised and still be large enough that closing it would move the very price it is measured against, and large enough that one participant's difficulty stops being that participant's difficulty alone. A ceiling is the instrument aimed at that. Nothing in the collateral machinery is aimed at it, and no amount of extra collateral turns into an opinion about size.

What does each rupee figure stand for?

Every rupee figure below answers one of three questions, and the question is worth asking as each figure arrives rather than afterwards. Did somebody agree it, quote it or work it out? Then it is a PRICE, and Rs 2,000.00/- is where the reference asset stands. Did it leave one account and reach another? Then it is a PAYMENT, and Rs 80.00/- is what an ordinary adverse move takes off a single unit. Or is it what survived after obligations pointing in both directions were set against each other? Then it is a NET, and no figure below is a NET. Only one of the two ever moves when the market moves, so a payment read as though it were a price makes the arithmetic come apart.

Quantities carry the same discipline, and it is worth stating separately because the two mistakes are not the same mistake. The units multiplied by the price give EXPOSURE. Exposure is the Rs 2,000.00/- a unit that the collateral is struck against, and the figure that decides what a move costs. A face amount multiplied by nothing that ever changes hands gives a NOTIONAL. A notional is a multiplier and never an amount at risk. Every quantity below is an exposure.

What is a position limit, in plain terms?

A position limit is an upper bound on the quantity of one contract any single participant is permitted to hold. The definition contains nothing else: no rate, no period, no formula and no threshold. A bound is a number somebody else has chosen and published. The figure itself is read from that authority.

No single ceiling covers the whole market in any case. One contract carries a different one from the next. Each of them is set by an exchange working inside a frameworkThe body of requirements an authority writes and revises for a market, inside which each exchange then settles its own detail. that the Securities and Exchange Board of India (SEBI) writes, and revisions arrive when the exchange judges they are needed, on no schedule that can be anticipated. The last column of the drawing below is empty in every row.

Why does collateral not already stop a position getting large?

Worked at two sizes, what refuses to change becomes visible. The reference asset here stands at a spot price of Rs 2,000.00/-. The same Rs 2,000.00/- is the EXPOSURE one unit carries. Anybody sitting on the reference asset collects nothing from it between now and the end of the period. The Rs 2,000.00/- is doing double duty below, as the price and as the whole of what a unit is worth to the person holding it. A payment arriving partway through would prise the two apart.

Against that unit sits collateral of Rs 160.00/-, an initial margin of 8.0 per cent of the exposure. Real margin percentages are set by clearing corporations working inside SEBI's requirements; 8.0 per cent is a round figure chosen here to keep the arithmetic legible. The Rs 2,000.00/- of exposure divided by the Rs 160.00/- put up gives 12.50 times. Take 4.0 per cent off the exposure, a perfectly ordinary day, and the PAYMENT is Rs 80.00/-. That payment takes 50.0 per cent of the collateral with it. A four per cent slip in the referenced thing takes half of what was lodged. Read that alongside the 12.50 times rather than after it: the multiple on its own sounds abstract, and the payment on its own sounds small.

Now take an invented twenty five unit holding. Exposure of Rs 50,000.00/-, collateral of Rs 4,000.00/- at the same invented 8.0 per cent, and Rs 50,000.00/- divided by Rs 4,000.00/- is 12.50 times again. The 4.0 per cent adverse move arrives as a PAYMENT of Rs 2,000.00/-, and it strips 50.0 per cent off the Rs 4,000.00/- lodged behind the holding. Read the two columns of the table below side by side and look for the reading that changed.

Nothing changed in the bottom two rows, and that is the finding a ceiling exists to answer. The exposure went up twenty five fold and the collateral went up twenty five fold with it, so every ratio the collateral machinery produces landed back exactly where it started. None of those ratios can see the size at all, so there is no size at which one of them turns round and says the holding has become too large. Collateral scales with the holding and therefore never becomes a reason to stop.

What is a ceiling actually for, if every position is collateralised anyway?

There are two reasons, and a reader handed only one of them walks away believing a ceiling is a form of protection wrapped around the person holding the position. A ceiling is nothing of the kind, and the two reasons make that plain by facing outward rather than inward.

The first reason is about the exit. A holding that is large in relation to the thing it references cannot be closed quietly. Somebody has to be on the other side of every unit that comes out, and if there are not enough of them at the price on the screen, the price moves to find them. The moving price is the same price at which every other position in the contract is revalued at the end of the day, so it is not a private matter between the holding and its buyer. One participant's exit becomes everybody's settlement price, which is a genuinely strange thing to meet for the first time. Think of a wedding hall that has taken a single booking covering four hundred guests. If that booking is withdrawn a fortnight out, the hall does not simply lose one customer. The hall starts offering dates at whatever price will fill them, and every other booking on its book was priced against a market that no longer exists.

The second reason is about the promise. ConcentrationWhen most of one contract has ended up in a single pair of hands, whatever the size of the market around it. is the word for one participant holding a great deal of one contract, and what makes it awkward is not the participant's own comfort. Whatever cannot be collected from one party still has to be met, and the somewhere it gets met from is shared with everybody else in the arrangement. So a promise big enough stops being one party's problem. Neither reason is about the reader, and giving only the first would quietly imply that it was.

What is a ceiling struck against, and why does that decide everything?

Here is the part a reader can actually put to work. A ceiling is always struck against something, and the something has a name: the base. A ceiling can be measured as a count of contracts. A ceiling can be measured as a share of what is outstandingThe running tally of positions in a contract that nobody has closed yet. The tally changes through the day as positions open and close. in that contract. A ceiling can be measured as a quantity of the referenced thing itself. Count, share and quantity are three different measuring sticks, and a holding that is comfortably inside one of them can be nowhere near comfortable inside another.

The same words describe completely different bounds depending on which base is meant, so a figure quoted without its base cannot be checked, cannot be complied with and cannot be compared with anything. Suppose the number quoted is two thousand. Two thousand contracts is a whole number to be counted. Two thousand as a share of what is open is not a fixed quantity at all. The denominator moves through the day, and the same holding drifts closer to the bound or further from it while its holder does nothing. Two thousand units of the referenced thing is a quantity that translates into contracts only once it is known how much of the thing one contract stands for. One numeral, three constraintsA bound on what may be done. A constraint takes choices away and hands nothing to the person it applies to. that have nothing in common but their spelling.

A base takes one of those three shapes. Its unit and its value belong to the authority named in the block above, and both move. The next question is what the shapes are for, and the drawing below sets one holding against all three.

When does a ceiling bite, and what happens then?

Most people first picture a ceiling as a fee, or as a penalty applied after the event. A ceiling is neither, and it bites in two places instead. A ceiling bites on the way in, in that a position may not be taken beyond it. And it bites on staying there, in that a holding which has drifted past the bound has to be brought back, and the drifting may have happened while its holder did nothing at all, if the base was a share of something that moves.

The second half surprises people, and is worth pausing on. A holding measured against a count sits where it was put. A holding measured against a share moves whenever the denominator moves. Positions elsewhere in the contract close, the open total shrinks, and a holding that was well inside the bound in the morning is pressed up against it by the afternoon without a single unit having been added to it.

A holder near a ceiling loses options rather than money, and losing options is exactly what makes this a decision rather than a cost. Nothing has been charged. Nothing has been taken. The ability to do a particular thing in a particular contract has gone. SEBI's requirements at sebi.gov.in settle what follows when a bound is passed, and how quickly it has to be put right.

What does a position holder near a ceiling actually have to decide?

The narrowness of the decision is what people miss, so the shape is worth naming precisely. A holder at the bound can add nothing further to that contract. So everything the position was still meant to achieve from here has to be achieved another way or left unachieved. The branch has no third arm. Notice what is not in it: there is no version where more of this contract is available at a price.

Suppose the position was taken out as a hedgeA position taken to offset an exposure somebody already carries, rather than to open a new one. How a hedge is designed for a business is covered separately. against something the holder already carries, and half of that job is covered. The bound does not care what the job was. The bound caps the contract, and the remaining half now has to be met from somewhere the bound does not reach, or accepted as uncovered. Compare a household with one salary paying for a school fee, a rent and a loan instalment. If the bank caps the standing instruction on that account, the household has not lost money that day. The household has lost a way of paying, and now has to find another one or drop a payment.

Naming what the decision is made of is not the same as making it. Five facts sit underneath any version of it. The job the position was taken on to do in the first place. How much of that job is already covered by something else already held. The base the bound is struck against, and where the holding sits against it. The collateral that has gone in so far. And what would have to be found if the price ran the wrong way while the holding sat there at the bound.

Who fills in which field, and which field can nobody fill?

The quickest way to see what a ceiling is worth in practice is to look at four different forms with a blank on each, and ask who fills that blank and from where. The forms belong to four different pairs of hands and only three of the blanks can be filled at all.

On a clearing member's own control sheet there is a field for utilisation: how much of the permitted amount the participants it acts for have used up. The risk desk fills it from two numbers it already has, the published bound for that contract and the holdings sitting on its own books read in aggregateA total taken across many separate holdings at once, rather than a figure read off any single one of them., and it fills it several times a day rather than once. Nothing about that field is a judgement. The field is a division.

On a surveillance analyst's sheet there is a field for how much of one contract sits in one pair of hands. The open total for the contract is the only thing that field can be filled from. Unless the denominator is stated beside it, the field is meaningless, and the base matters for exactly that reason. The analyst is not asking whether the holder is comfortable. The analyst is asking whether an exit by that holder would move the price everybody else is revalued at.

On a lender's schedule, where a position has been taken as security against an advance, there is a field for what the security could be turned into and how fast. The published bound helps in an indirect way. It tells the lender the outer edge of what any single party can be carrying in that contract, and that outer edge caps how bad the crowding at the exit can get. The bound says nothing whatever about this borrower.

And on a household's own sheet of paper there is a field that none of the three forms above can fill: how large a position would suit the household. The member's field is a division. The analyst's field is a share. The lender's field is about security. Not one of them was ever computed with this household in mind. Knowing which fields are fillable, and from where, is more use to the household than a number that was never about it.

What does a ceiling not settle?

Three negatives, and each of them is a live misreading rather than a hypothetical one.

The first: a ceiling is not a statement that a holding beneath it is prudent. The ceiling was struck for the arrangement as a whole, and no individual was in anybody's mind when it was set. A speed limit on a road was not calculated from the condition of any particular driver's tyres.

The second: a ceiling is not a target. A holder who reads the bound as the size they are allowed to reach, and works towards it, has turned a bound into an entitlementSomething a person is owed and may claim. A rule that sets a ceiling grants nobody anything. It only takes options away., which is a category error rather than an aggressive choice. Nothing was granted. Something was withheld.

The third: a ceiling says nothing whatever about whether the holder could meet a call on the position they already hold. On the invented figures above, a holding at any size under any bound still gives up 50.0 per cent of what was lodged on a 4.0 per cent adverse move, at 12.50 times of exposure for each rupee lodged. The ratio does not know how large the holding is, so no bound placed on the holding can change it.

Why is the holding here a size rather than a share of what is open?

A share needs a denominator, and the denominator under a position limit is a count of how much of that contract is open. An exchange publishes such a count for a contract that trades. The reference asset worked through above has no such count behind it, so the twenty five unit holding can only be a size: there is no total for it to be a fraction of, and therefore no distance to any bound.

A control walking a holding towards a bound would have had to invent two things at once, the bound and the total the bound is measured against. A reader who dragged that control would come away believing they had watched a cap bite. They would in fact have watched two made-up figures approach each other at a made-up speed. The arithmetic above shows what a ceiling does. Showing a holding approach one takes a manufactured count sitting underneath, and a manufactured count teaches nobody where a real bound sits.

One more thing in the arithmetic looks like a slip and is not. The 4.0 per cent adverse move on the twenty five unit holding arrives as a PAYMENT of Rs 2,000.00/-. The reference asset stands at Rs 2,000.00/- as a PRICE. Neither figure was copied into the other. Twenty five multiplied by four per cent comes to exactly one, so the move on that holding is worth precisely one unit of exposure, and the two land on the same numeral by force of the arithmetic rather than by chance. Remember also that no payment reaches a holder of the reference asset at any point in this, so the Rs 2,000.00/- price is the entire worth of a unit to whoever holds it.

So how large a position should anybody carry?

Anybody who has followed the argument this far arrives at exactly that question, and it is owed something flat rather than something careful. Put plainly: nobody can answer it for somebody else. A ceiling is not a stand-in answer either. A ceiling was struck for the arrangement and not for any one holder, which is the point the failure block above makes in a different direction.

Settling it would take five facts, and they are worth naming in the order somebody would actually meet them. The job the position was taken on to do. How much of that job is already covered by something else already held. The collateral that has gone in so far. The sum that could be raised at a day's notice if the price ran the wrong way for a week. And what actually lands in the account on the worst morning rather than an ordinary one. Every one of the five is a fact about one holder on one day, and not one of them can be read off a contract specification, a rule book or a published bound. Only the person holding all five can settle the question. Nobody else has the facts to settle it for them.

Understanding how a ceiling operates is a reading skill, and it is not a reason to be standing under one. The parts the decision is made of, met in the order somebody would actually meet them, are the useful thing anybody can be handed.