Open Interest Data: What It Shows and What It Cannot

Open interest counts the contracts still outstanding at a moment. Every contract has a long side and a short side, so open interest counts each contract once rather than twice, and a rise of one contract means one new long side and one new short side arriving together. Open interest shows how many promises remain open, and shows nothing about who wanted them or what follows.

Open interest is taught here with no level to look at: no number, no series running back across a week, no chart with a line climbing across it. Working without a level is the demonstration itself. The moment a plausible looking level is printed for open interestThe number of contracts still outstanding at a moment, counted once for each contract rather than once for each side., a reader starts studying the level and stops studying the measurement, and the measurement is what has to be got right first.

Here is the situation, and it is a common one. A line arrives in a note, in a message, or in a caption printed under a chart. The line says that open interest rose, and that the rise shows the market leaning to one side. The arithmetic in the note is fine. The tone is calm and technical. Nothing about it reads as careless. And the sentence still cannot be true, not because somebody used poor data, but because of what a count of contracts is. The mistake is one of the very few errors in this subject that can be settled completely rather than argued over, and settling it takes about a minute.

Everything below rests on one plain observation. A count is a count. There is no direction inside it, there is no price inside it, and there is no money inside it. Every honest use of this figure comes from noticing which of those three a sentence has quietly assumed it already has. Turning the count into money needs one further number that no account of the subject can supply from memory. Reading a direction out of it needs a number that no count contains. A count answers how many, and every question that is not how many needs something else carried to it from somewhere else.

What does open interest count, exactly?

Open interest is the number of contracts still outstanding at a moment: promises that have been made and not yet closed outEnded by taking the opposite position rather than by waiting and settling at expiry., settled or expired. The definition ends there.

Now the distinction that catches almost everybody, and it deserves a sentence of its own. A count of how many contracts changed hands during a day and a count of how many remain open at the end of it are two different measurements, and the second can fall on a day when the first is enormous. The two counts are not two views of the same quantity. Nor are they a fast version and a slow version of one number. The two counts answer different questions, and a day of furious activity in which most of what was written was also unwound will show a large one and a shrinking other.

The everyday version makes it obvious in about four seconds. Think of a restaurant on a Saturday. The owner can count two quite different things. She can count how many parties came through the door since she opened, which climbs all evening and never falls. Or she can count how many tables are occupied right now, at nine in the evening, which goes up when a party is seated and down when one pays and leaves. Both are honest counts of the same evening. Only a confused manager would use one where the other was wanted. Open interest is the tables occupied at nine. The count of contracts that changed hands is the parties who came and went.

Hold on to the fact that the second count can fall while the first is large. Confusing the two is the source of a great deal of confident nonsense. A busy day is not the same thing as a growing pile of open promises. A busy day can be one on which a very large number of existing promises were unwound. Unwinding on that scale takes an enormous amount of dealing and leaves fewer contracts open at the end than there were at the start.

Why does every change in the count arrive in pairs?

One mechanism sits underneath everything else here, and hardly anybody is ever handed it. A contract cannot exist with one side. A contract is an agreement, and an agreement with nobody on the other end of it is a wish. So for open interest to rise by one contract, a long sideThe side of a contract that gains as the referenced price rises. and a short sideThe side of a contract that gains as the referenced price falls. have to arrive at the same moment, on the same terms, and agree.

Everything that follows leans on the consequence, so it is worth taking slowly. An increase in open interest is one more of each side, so an increase cannot by itself mean that more people want one side than the other. Not usually cannot. Cannot at all, in the way that a triangle cannot have four corners. There is no arrangement of the arithmetic in which one side outnumbers the other. What is counted is the agreement, and an agreement that only one party wanted did not happen.

There is a second half to this and it explains the shape of the number. The pair is counted once, not twice, and that is why open interest is a count of contracts rather than a count of positions. One new long side and one new short side arrive, and the figure moves by one. If the convention were the other way, the same event would move it by two, and the number would be a count of the people standing rather than a count of the promises standing. Knowing which convention is in use is worth something, and this one counts the promise.

The household version: a marriage register does not record two marriages when two people marry. The register records one, with two parties to it. Nobody looks at a register that grew by one entry and concludes that more people wanted to be one half of a couple than the other half. The absurdity is obvious the moment the counted thing is something easy to picture, and it is exactly as absurd when the counted thing is a contract.

What does open interest actually show?

Open interest can start to look useless after so much about what it cannot say. The count is not useless at all. The count shows how many promises are open, a statement about outstanding obligation and about nothing else.

Open interest is genuinely good for one thing: it states how much is still to be settled or closed out. How busy a day was is a different question, and where anything is priced is a different question again. The three questions get muddled constantly, and keeping them apart is most of the skill. How busy was it is answered by the count of contracts that changed hands. Where is it priced is answered by a price. How much remains to be dealt with is answered by open interest, and by nothing else on the screen.

A courier depot at the end of a shift works the same way. The number of parcels that passed over the counter today shows how hard everybody worked. The price on the tariff card gives what a delivery costs. Neither gives the number of parcels still sitting in the building waiting to go out, and that third number is the one that decides whether tomorrow morning is calm or unpleasant. Open interest is the parcels still in the building.

Now the honest limit. Knowing how many promises are open gives the size of what is outstanding and nothing whatsoever about its composition. The parcels in the depot could be one enormous order for a single street or a thousand separate addresses across the district, and the count is identical. The count does not give the price at which each of those promises was written, the day, the party, or the reason. All of that is inside the number in the sense that it happened, and none of it is inside the number in the sense that it could be got back out.

What can the count not show about who wanted the contracts?

The count carries no record of which side approached the other, which side was more eager to have the thing written, or which side gave ground on price to get it done. None of that is missing because somebody failed to collect it. The measurement never had anywhere to put it.

The figure counts completed agreements, and a completed agreement has two willing sides by definition, so nothing about the negotiation survives into it. Everything interesting about who wanted what happened before the moment the count cares about. One side may have chased for a week. One side may have shifted on terms at the last minute. One may have been indifferent and agreed because the terms happened to be acceptable. All three histories produce the same entry: a contract, outstanding, counted once.

Here is the everyday form and it is worth keeping. A figure that counts agreements says nothing about disagreement. A register of completed house sales in a locality gives the number of houses that changed hands. The register does not say which sellers held out for months and which took the first offer, and no amount of staring at it will ever produce that. The register was never a record of the haggling, only of the handshakes.

The point matters more than it sounds. The sentence people most want to write about open interest is a sentence about eagerness. Somebody wanted these contracts, and surely a rise names who. It does not. A rise says only that a pile of promises got larger, and both sides of every one of those promises walked in willingly.

What can the count not show about what happens next?

A count taken at a moment is a statement about that moment. A count is not a forecast wearing a modest hat, and it does not become one when several such counts are put beside each other.

To say that one direction has become more likely requires a likelihood attached to a price. A count of contracts attaches no likelihood to any price at all. So that sentence is not available, and nothing is substituted for it: no estimate, no rough feel, no in general such rises tend to be followed by. A count holds no series, no distribution, no probability and no recorded outcome, and a sentence that quietly supplies one has invented it and dressed the invention as a reading.

Something is available, and it is worth saying precisely: the number of promises outstanding has changed, and whatever was true about the obligation those promises carry is now true of more of them, or of fewer. The change is a real fact about the world. If every one of those contracts obliges somebody to deliver something at a stated price, and there are more of them open tonight than there were last night, then there is more of that obligation standing about. The change says nothing about which way anything moves next. It says something true about how much is outstanding.

The reason people slide from the first to the second is that a rising number feels like momentum. Numbers that go up feel directional. But the count of open contracts goes up when two willing sides write a fresh promise, and the count of tables occupied at nine goes up when a party is seated. Neither says anything about the rest of the evening.

How is a count of contracts turned into an amount of money?

Here is a check worked rather than described. The number the sentence would need in order to be true is named, and then looked for. If it is there, the sentence can be written. If it is not, the sentence is not available and something shorter and true is.

The sentence worth writing is: this many open contracts represents this much money. Three things are needed to write it. The first is a count of open contracts, which is the reading in front of the analyst. The second is the contract sizeThe number of units of the underlying that one contract stands for, which turns a count of contracts into a number of units. in units, which is what turns one contract into a quantity of the underlying. The third is a price for the underlying, and this guide has one: the invented reference asset is at Rs 2,000.00/- a unit, and it pays nothing at all while it is held.

Two of the three are available and the middle one is not. The contract size in units is set by the Securities and Exchange Board of India (SEBI) at sebi.gov.in, it varies from one contract to another, and it moves. Writing a number in there from memory would be inventing a specification. A hole announces itself and an invention does not, so the invention is the worse fault.

So the multiplication is written with the middle term left as a labelled blank, and the authority that fills it is named. The count of open contracts, multiplied by the contract size in units, multiplied by the price of Rs 2,000.00/- a unit, gives an amount of exposureThe amount of the underlying a position actually stands against, as distinct from what was paid or posted.. Two terms written, one named and empty.

Now the version worth carrying into everything else. An incomplete multiplication with a named blank in it is worth more to a reader than a completed one with a guessed term sitting in the middle. The first states exactly what is not yet known and where to go and get it. The second states a number, with no way of telling which part of it was real. A household comparing two loan offers meets the same thing: an instalment figure with the tenure left off is a blank that can be filled by asking, and an instalment figure with the tenure quietly assumed is a wrong answer that looks finished.

What is posted behind all those open contracts?

The same missing figure turns up one step further along, at exactly the place a reader will next want a number. Every one of those outstanding promises has something posted against it while it runs. The posted amount is called marginThe amount posted against an open position while that position remains open., and margin is what makes these contracts different to hold from anything bought outright.

Worked for one unit: at an initial margin of 8.0 per cent of the exposure, invented for teaching, Rs 160.00/- is posted against Rs 2,000.00/- of exposure. The two figures are the same number in a different role. The exposure standing behind one unit is the price of one unit, Rs 2,000.00/-, so a price and an exposure agree here by construction rather than by coincidence.

The 8.0 per cent gives a figure per unit, and that figure still cannot be carried up to a count of contracts without the contract size. The contract size is the identical term that was missing a moment ago. Multiplying Rs 160.00/- by a count of contracts quietly assumes that one contract is one unit, and there is no reason on earth for that to be true. The same hole appears in the same place, one row lower.

The 8.0 per cent of the exposure worked with here is a teaching figure, not a requirement anybody set. Real margin is worked out by clearing corporations operating under SEBI's framework at sebi.gov.in, it varies by contract and by day, and it moves. The 8.0 per cent is not a fact about anything and carries no authority.

What would actually change hands under all those contracts?

A count of open contracts is a headline quantity like any other, so one question works on it as on every headline quantity. What would actually change hands, and over what period? Then divide the headline by that sum and look at the answer.

The clearest illustration of the gap in this record is not a count at all. The illustration is the two legged arrangement between the fixed payer and the floating payer, and it shows the distance at its very widest. A notionalThe amount that the payments under an arrangement are multiplied by, which is never itself paid across. of Rs 1,000 crore against a first period net of Rs 12.00 crore makes the notional 83.33 times the sum that actually moves. Do both divisions rather than one, because they teach different halves of the same fact. Rs 1,000 crore divided by Rs 12.00 crore is 83.33 times. Rs 12.00 crore divided by Rs 1,000 crore is 1.2 per cent of the notional, which leaves 98.8 per cent of that headline sitting exactly where it was.

The notional never changes hands. The notional is a multiplier and not an amount at risk, and an account that reports the headline without saying so has told a reader the arrangement is something like eighty three times larger than the cash it produces in a period.

Here is the general form, and it is the most portable idea here: a number that measures how much is outstanding is not a number that measures how much is moving, and reaching for the first when the second was wanted is one habit away from being avoided. A count of open contracts stands in exactly that relation to whatever settles under them. A count measures the size of the pile of promises. It does not measure the sums that pass in a period, and no amount of care in reading the count will convert one into the other.

The household version is a housing society. The total value of every flat in it is a headline quantity, and it is enormous. The maintenance collected this month is what moves. Both are true numbers about the same society and only one of them says anything about this month, and a resident who mixed them up would either panic or relax for entirely wrong reasons.

What can be said honestly with no level in front of the reader?

A count of contracts holds no history of any kind, so no level, no series and no trend comes with it. Holding no history turns out to permit a great deal rather than very little. Most readers expect the opposite, and the reversal is the most portable idea here.

Three sentences remain available and each one is worth having. First: a change of one contract is one new long side and one new short side. Second: a count of outstanding contracts and a count of contracts traded are different measurements. Third: a count says nothing at all about which side approached the other. Every one of those is true today, was true last year, and will be true of any market in any country, because each follows from what a contract is rather than from what any reading happened to be.

One kind of sentence is not available and it is worth naming precisely: anything beginning with the words the level is, or the trend has been, needs a series this record does not hold. The limit is not modesty and it is not a hedge. The limit is the difference between what follows from the structure of a contract and what would need an observation. The first travels anywhere. The second has to be fetched, and fetched from a named source, on a stated date.

Sorting sentences that way makes a lot of writing easy to read. A claim either follows from what a contract is, or it needs an observation somebody made. Both kinds can be true. Only the second kind can be out of date, and only the second kind can be quietly missing its source.

What does reading this figure carefully not do?

A reader who has just been shown three things a number cannot say will quite reasonably feel better equipped, and the honest thing is to say exactly what they are better equipped at.

A reader is better equipped at reading. That is the whole of it, and it is worth having. Nothing about what any contract obliges has changed, nothing about what is owed at any price has changed, and correcting a misreading is not a form of protection. If a contract obliged somebody to deliver at a stated price this morning, it obliges exactly the same thing this evening, whatever anybody now understands about the count of how many such contracts are open. A check catches a misreading. A check does not reach into the obligation and soften it.

Now the part that matters most and gets written least. Somebody who read a rise in open interest as a lean towards one side, and acted on that reading, was reading the figure the way it is described in a great deal of print, including print produced by careful people. The phrasing has been in circulation long enough to sound like a fact rather than a claim. Treating a widely repeated explanation as carelessness teaches blame rather than reading, and blame is not a skill anybody can use.

So the correction here is aimed at the sentence, not at the person who believed it. Anybody who has written that sentence, or repeated it, or nodded along at it in a meeting, is in the company of most people who work with these figures. The change is in what happens the next time the sentence appears.

How does somebody who never deals in these contracts use any of this?

Most people who need this distinction will never be a party to one of these contracts in their lives. Most people meet the figure secondhand, in a caption, in a message forwarded by a relative, in a paragraph of a market wrap read on a phone at the end of a long day. Secondhand reading is the setting that matters most, and it carries four quite different uses.

A household is the first and the most important. A message arrives saying that open interest in something rose sharply, and that this shows where the smart money is going. The household has no way to check the figure and no reason to. In four seconds the household can ask what the other side of every one of those contracts did, and the sentence falls over on its own, without any data, without any expertise and without any argument about the market. The question is a defence that costs nothing and needs nothing fetched.

An analyst uses it as a filter on writing rather than on markets. A note that reads a rise as a lean has said something about the writer's care, and it has said it before the reader reaches the parts of the note that cannot be checked. A note like that is cheap evidence about expensive claims, and cheap evidence about expensive claims is the most valuable kind.

A lender looking at a borrower who deals in these contracts wants the third use. The count of open contracts sitting against a borrower is a measure of how much is outstanding, not of what settles in a period, and those two questions need entirely different numbers. A lender who reads the first and thinks about the second has misjudged the size of the thing in front of them, in whichever direction their assumption happened to lean.

The fourth use is the one that outlasts this subject entirely: any headline quantity, anywhere, invites one question: what would actually change hands in a period, and the answer is very often a small fraction of the headline. An investor reading about a corporate arrangement, a journalist reading a press release, a resident reading a society notice. The habit was learned here on a count of contracts and it does not stay here.

Who sets the things not printed here?

Four requirements are touched by everything above. The authority is named for each of them and the value for none, and a fifth row holds the one invented percentage used here so that it can never be mistaken for something an authority set.

The value cell in each of the four rows is empty. Every one of those four is set by the authority printed inside the row, and every one of them moves. A figure written out here would be wrong rather than merely old, and there would be no way to tell which, because a stale number and a fresh one look exactly alike on a screen. The fifth row holds the 8.0 per cent, which is worked with here and marked as an invention so that it is never lifted out and repeated as a requirement.

Should a reader who can read this figure enter into one of these contracts?

The question arrives here, so it is answered here. Whether anybody should be a party to one of these contracts turns on facts about that person, and no general account holds them.

Naming those facts is the useful thing. A refusal that names what is missing is worth something, and a refusal that just says no is not.

A count of open contracts supplies not one of those four, so reading it correctly leaves the question exactly where it was. Reading it correctly leaves the reader better at reading a sentence somebody wrote about a figure, and being better at that is genuinely worth having. The third row is out of reach here. A count of contracts attaches no likelihood to any price. The fourth is out of reach anywhere. Everything in it depends on who is asking.

The reference asset, its price and the two legged arrangement between a fixed payer and a floating payer are invented.
Educational material. Not advice on any investment, tax, budget or market position.