Event-Driven Strategies: Trading Around a Corporate Event

Event-driven is approach 5 of the eight approaches covered in this series. The position is taken because a specific corporate event is expected to complete, so what decides the outcome is the event rather than the market. The gain if it completes is the small difference between the price paid and the offer, and the loss if it fails is the whole way back.

Start with the one structural feature that separates this approach from most of the others beside it. A position taken around an announced corporate event has an ending written into it before the fund has held it for a single day. Something has been announced. The announcement will either be carried out or it will not, and on the day the question is settled the position stops being a position: either the offer is paid and the shares are gone, or the offer collapses and there is nothing left to wait for. The fund does not choose the date and does not choose the answer. The fund has bought its way into somebody else's decision and is now waiting for that decision to arrive.

Hold that against the shape of a position that depends on a price. If a fund holds shares because it thinks they are worth more than they cost, nothing in the world will ever tell it that the question has been settled. There is no day on which the market rings a bell. The position ends when the manager decides to end it. The exit is a judgement rather than an event. Several of the eight approaches covered in this series work that way, and this one does not, and that difference is worth carrying through everything below.

What does it actually mean to say a position depends on an event?

At household scale the shape is easier to see, so start with a smaller case. A household owns a scooter it was going to sell for around Rs 60,000. A neighbour comes and says, in front of witnesses, that he will buy it for Rs 60,000 and will hand over the money on Friday when his own payment arrives. The same afternoon a second person offers Rs 57,000 in cash, right now, no waiting.

The Rs 3,000 between those two numbers is not an opinion about the scooter's value. Both people have already agreed it is worth about Rs 60,000. The Rs 3,000 is the price of Friday: of waiting four days, and of the chance that Friday comes and the payment has not. Taking the Rs 57,000 sells a scooter. Refusing it and waiting means no longer holding a scooter at all. Refusing it means holding a position in whether a particular neighbour keeps a particular promise on a particular day.

The scooter case is the whole of it. Event-drivenA position whose outcome depends on a specific corporate event rather than on a market move. investing is the same situation at the scale of listed companies, where the promise is a corporate transaction that has been announced and the four days are however long it takes for that transaction to be settled one way or the other. The question the position actually turns on is whether the announced thing happens. The market can rise or fall in the meantime and it changes nothing about that question.

The announcement changes something structural rather than something of degree, and the change is worth stating precisely. Before the announcement, the shares of a company have one number attached to them: whatever the market says today. After an announcement, they have two. There is still the traded price, and beside it there is now a stated price that somebody has said they will pay. The traded price and the stated price are not the same and are not meant to be. The position lives entirely in the distance between them.

One constructed case runs from beginning to end below, and the case is built rather than borrowed. Nilgiri Absolute Return Fund, invented, is the vehicle these hedge fund materials describe, and it runs approach 6 rather than this one. At the record date its long positions were Rs 6,50,00,00,000 and its short positions Rs 2,50,00,00,000 against net assets of Rs 5,00,00,00,000, which is gross exposure of 180.0 per cent and net exposure of 80.0 per cent. The position worked below is one a fund could hold under approach 5, and it is not a position the Nilgiri fund holds.

How does an announced offer put a second number beside the price?

How Event-Driven Strategies Work

Here is the constructed case, and it is only three numbers. A company's shares were trading at Rs 80.00 and nothing unusual was happening. Then a transaction is announced, and the announcement carries a stated price of Rs 100.00 a share. Within the day the shares are trading at Rs 96.00. Those three numbers are the whole set-up, and every number that follows is worked out of them.

The Rs 80.00 is the undisturbed priceThe price the shares traded at before anything was announced.: what the shares were worth to the market when the only thing anybody was pricing was the company itself. The Rs 100.00 is the announced offerA stated price somebody has said they will pay for the shares.: a price somebody has stated they will pay. A stated price is a fact about a document rather than an opinion about value. The Rs 96.00 is where the two meet in a live market: high because the offer exists, and short of Rs 100.00 because the offer has not been paid yet.

An event-driven position is the deliberate purchase of the distance between the traded price and the announced offer. The fund buys at Rs 96.00. If the transaction completes, Rs 100.00 a share arrives. The distance is Rs 4.00. The distance has a name worth using carefully: the gapThe distance between the traded price and the announced offer, which is what the position earns if the event completes.. Everything a reader gets wrong about this approach comes from misreading what that one number is and what it is not.

Everything that follows turns on getting the denominator right, so here is one more term before the arithmetic. The fund pays Rs 96.00 a share. The Rs 96.00 paid is the money at riskThe amount actually paid for the shares, which is the denominator both outcomes are measured against., and every percentage below is struck on it. Not on the Rs 100.00 that was offered, not on the Rs 20.00 between the undisturbed price and the offer, and not on the Rs 4.00 gap. The money that left the fund is Rs 96.00 a share, so the money that can be lost or gained is measured against Rs 96.00 a share.

Why do the shares sit at Rs 96.00 rather than at Rs 100.00?

Most readers skip this question, and skipping it is why the gap gets misread later. Somebody has stated they will pay Rs 100.00. If that were the end of the matter the shares would trade at Rs 100.00. Anybody buying at Rs 99.00 would be collecting a rupee for nothing, and the buying would push the price up until there was nothing left to collect. The shares sit at Rs 96.00 instead, and the Rs 4.00 that is missing is not an accident or an inefficiency. The missing Rs 4.00 is a price, and it is being charged for two separate things at the same time.

The gap is one number doing two jobs, and a reader who treats it as one thing has understood neither. The first job is time. Rs 100.00 that arrives after a wait is not the same thing as Rs 100.00 that arrives today, and that is true even if the arrival is completely certain. A fund that has borrowed to hold the position is paying for the wait every day of it; a fund that has not borrowed has still tied up rupees that could have been somewhere else. Discounting a future amount is covered separately, and the same machinery applies here to a payment whose date is unknown.

The second job is completionThe event actually happening, which is the moment the offer price is received. itself. An announcement is a statement of intent by parties who may not be able to carry it out. Boards change position, shareholders vote things down, lenders withhold consent, and a clearance that a transaction needs may not arrive. The price is short of the offer partly because the offer may never be paid at all, and the market has quietly put a number on that when it settled at Rs 96.00.

Now the important part, and it is a limit rather than a technique. The Rs 4.00 is not split into a part for the wait and a part for the doubt. Splitting it honestly would take two inputs: a rate at which waiting is priced, and a likelihood that the transaction completes. The second is a judgement about a specific transaction rather than arithmetic, and without it the split is a guess dressed as arithmetic. So the gap holds both, the direction each pushes is stated, and the dividing line is left where it falls. The lesson is not a decomposition. The lesson is that the single most visible number on the screen is a compound, and reading it as a simple one is the first mistake available.

Which kinds of corporate event does this cover?

The worked case uses an offer for a company's shares because it is the cleanest one to draw, but the approach is not limited to that. All of these share the structural feature set out at the start: something has been announced, and there is a day on which the announcement is either carried out or abandoned. How each of these transactions actually works inside the company is covered separately; here they matter only as the thing the position waits for.

Look down the third column, the same column five times. Every row ends in a binary, and every binary has a date attached to it even where nobody yet knows what that date is. The repeated shape, and not the corporate detail of any one row, is what makes these events tradeable as a single approach. A fund running this approach is not an expert in demergers or an expert in buybacks. The fund is running a book of positions whose common feature is that each of them stops existing when a question somebody else is answering finally gets answered.

The last row is worth a moment because it is the one that fits least comfortably. A change to a reference index has no bidder and nobody is offering to pay anything. The index change has instead a stated date on which a large, mechanical amount of buying or selling is expected to happen, and a stated date is enough to give a position a defined ending in exactly the same sense. The reference broad equity index used in these worked cases is a constructed one.

What does the position gain if it completes, and what does it lose if it does not?

Take the completion first, the easy half. The fund paid Rs 96.00 a share. The transaction completes and Rs 100.00 a share is received. The gain is Rs 4.00 a share, and against the Rs 96.00 that actually left the fund that is 4.17 per cent. Rs 4.00 is the entire upside. After completion the shares are gone, so there is no second payment, no revision upwards and no participation in anything that happens afterwards. The gain is capped at the announced offer by construction, and nothing that goes well can make it larger.

Now the failure, and this is the half most treatments skate over. Suppose the offer collapses. Not a delay, not a revision, but an abandonment: the transaction is off. Where does the price go? The price does not fall to something a little below Rs 96.00. The only reason it was ever at Rs 96.00 was the announcement, and the announcement has just been withdrawn, so the price goes back towards the Rs 80.00 the shares stood at before anybody said anything. The journey back down is the fall backThe distance from the traded price to where the shares stood before the announcement., and on these numbers it is Rs 16.00 a share, being 16.67 per cent of the money at risk.

Set them beside each other and look at what kind of quantity each one is. Rs 4.00 is a distance up to a stated offer, and it is small precisely because the offer is believed: if the market thought the offer were worthless the shares would not be anywhere near Rs 96.00, and if it thought completion were a formality the shares would be at Rs 100.00 and there would be no position to take. Rs 16.00 is a distance back to a price that was never about the transaction at all. The two numbers are not two sides of one bet; they are a small distance to a promise and a large distance back to the world without it.

At what price paid do the two outcomes become the same size?

Everything above used one price paid, Rs 96.00, so the one to four came out of it. But the fund does not have to buy at Rs 96.00, and different funds buying on different days after the same announcement will hold the same position at quite different prices. So the honest question is what the shape looks like across the whole range, from paying the old undisturbed price of Rs 80.00 to paying the full announced offer of Rs 100.00.

The arithmetic is short enough to do mentally. At a price paid of P, the gain if the transaction completes is Rs 100.00 less P, and the loss if it is abandoned is P less Rs 80.00. Add those two together and P disappears: they sum to Rs 20.00 at every price. Rs 20.00 is the whole distance between the undisturbed price and the offer, and the price paid merely decides how that distance is divided. One outcome grows exactly as fast as the other shrinks, so the two can only be the same size at one price, and that price is the midpoint.

The midpoint of Rs 80.00 and Rs 100.00 is Rs 90.00, and at Rs 90.00 the gain on completion is Rs 10.00 and the loss on abandonment is Rs 10.00. Above Rs 90.00 the loss is the larger of the two and the asymmetry grows the closer the price gets to the offer. Below Rs 90.00 the gain is the larger. None of that is a statement about which outcome arrives. The midpoint is a statement about how the same Rs 20.00 gets split, and it holds whatever anybody thinks of the transaction.

Does the same Rs 4.00 get any bigger if the wait gets longer?

No, and the flatness of that answer is the point. The gain is fixed by the announced offer. Rs 100.00 was stated, Rs 96.00 was paid, and Rs 4.00 is what the completion pays whether it arrives after one quarter or after four. Nothing about a long wait improves the amount. The rupees are settled the day the position is opened; only the date is open.

A longer wait changes how much time those same rupees occupy. A quarter is taken as 91 days throughout these worked cases, so one quarter is 91 days and four quarters are 364. The same Rs 4.00 spread across 364 days instead of 91 is the same money doing the same job over four times as much of the fund's life. Everything else the fund could have done with those rupees was not done, and if the position was held with borrowed money the borrowing was running the whole time.

Here is where the arithmetic stops. The obvious next move is to divide the Rs 4.00 by the time held and quote a rate a year, and exactly that is done in plenty of places. A rate a year needs a holding period, and the holding period is the one thing nobody knows on the day the position is opened. An annual figure built on a guessed date is an assertion about the date rather than about the transaction, and it reads as a promise the transaction never made.

Who is actually deciding whether this position resolves one way or the other?

Not the fund, and that is the risk the approach carries. Every other input the manager can work on. The manager can read the announcement, can read what each side has committed to, can watch the traded price move and can size the position. The manager cannot cast a vote it does not hold, give a consent it is not asked for or issue a clearance it does not issue. The fund has taken a position in a decision that belongs to other people, and no amount of further analysis converts that decision into something it controls.

Event-driven sits next to two approaches that look similar on the surface and are not, so two boundaries are worth drawing. A position taken because a difference between two prices is expected to close is relative value, and the mechanism there is the closing itself. Arbitrage in its precise sense is a difference that can be locked at the moment it is seen, with no further decision by anybody needed. The gap worked here is neither. The gap closes only if a decision goes one way, and it cannot be locked. Nobody can fix today the outcome of a vote that has not happened. Both of those mechanisms are covered separately.

Two more edges, stated once and then left alone. Some event positions carry a short leg alongside the shares held, and there is one thing to carry from that: a short position's gain is bounded and its loss is not, and how stock is borrowed and sold short is covered separately and is not re-worked here. And a corporate event traded in a liquid market is not the same subject as a private fund selling one of its holdings; the routes by which a private vehicle disposes of what it holds have their own treatment, and blurring the two would be a genuine error rather than a simplification.

What goes wrong when the gap is the most visible number on the screen?

The error worth naming here is not an error of arithmetic but an error of attention. Somebody describes the position the way it is usually described: bought at Rs 96.00 against an offer of Rs 100.00, a gap of Rs 4.00. Every number in that sentence is correct. A reader hears it, notices that Rs 4.00 is a small number, and concludes that not much is at stake. Small gain, small risk, and the position sounds modest.

The position is not modest. If the offer is abandoned, Rs 16.00 a share is at stake, four times the gain. The price does not fall back to a little under Rs 96.00. The price falls back towards the Rs 80.00 it stood at before anybody said anything. The gap and the fall back are not two versions of the same distance and they are not even measured from the same reference point, so sizing one by looking at the other is wrong by construction. The reader who does it has under-stated the loss by a factor of four on these numbers, and would under-state it by more at any price paid above Rs 96.00.

Which number cannot be supplied by arithmetic, and why is saying so the honest answer?

Everything above can now be computed. Gain Rs 4.00 if it completes. Loss Rs 16.00 if it does not. One to four. Crossing at Rs 90.00. One input is still missing, and it is the input everything turns on: how likely the announced transaction is to be carried out. Without that likelihood, whether the position is worth taking cannot be said.

The completion likelihood is not arithmetic. The likelihood is a judgement about specific boards, specific shareholders, specific lenders and specific clearances in a specific transaction, and a judgement of that kind has to be defended by whoever puts money behind it. Everything else here is arithmetic and comes straight out of the three numbers. The one thing that is not arithmetic is a judgement, and a range or a rough figure put in its place would do nothing but make the worked example feel complete.

The mechanism that remains is more than nothing. The shape of the two outcomes is now known, and what each is measured from, and that they always add to the same Rs 20.00, and where they cross, and that a long wait does not enlarge the gain. Knowing all of that is a real understanding of the mechanism. Knowing it is not a decision, and a completion likelihood offered in order to feel more useful would swap something true for something guessed.

How would somebody reading a fund factsheet actually use any of this?

Most readers will never take a position of this kind. A good many of them will read a document in which a manager describes what it does, and that is where the mechanism becomes practical. Suppose a factsheet or an offering document says the manager runs an event-driven approach. Four questions then follow that the document either answers or does not, and noticing which is the whole skill.

First, does the document state the price the downside is measured back to? A description that gives the gap and stops has given the smaller of the two numbers and left out the larger. The undisturbed price is the missing input, and its absence is not a detail. Second, is any gain in the document quoted as an amount or as a rate a year? Both can be honest, but a rate a year has had an assumption about the wait folded into it, and the reader is entitled to know which wait was assumed before comparing that figure with anything else.

Third, how many positions are running at once, and are the events independent of each other? A book of event positions can look diversified by name and still be a single position if the same conditions have to be satisfied for all of them. Counting positions is not the same as counting sources of risk, and this is the reading error that survives longest. Fourth, what does the document say happens when an event fails? A description that dwells only on completion has described half of a two-sided structure, and the half just worked through above is the bigger one.

There is a household version of exactly this, and it is worth carrying because it makes the shape stick. A household agrees to sell a plot of land. The buyer has signed, the paperwork is moving, and the money is expected in three months. On the strength of that, the household commits to a purchase that has to be paid for next month. The gain from waiting for the agreed price rather than accepting a lower cash offer today is a few per cent. If the sale falls through, the shortfall is not a few per cent of anything. The whole of the second commitment sits there with nothing behind it. Same shape, same asymmetry, same missing number, and nobody in the story ever needed to know what a hedge fund is.

The vocabulary around this approach invites one wrong idea, and it is worth closing on. A mechanism is one thing and a recommendation is another. Whether an event-driven position is better than any other, whether it suits a particular investor, whether it belongs in a portfolio and what it returns are four separate judgements, and every one of them needs the completion likelihood the arithmetic above cannot supply. The mechanism and the risk it carries are what the arithmetic gives, and the four judgements sit outside it.