The Leak: When a Transaction Becomes Public Too Early

A leak is a transaction becoming known outside the group holding it before it has been announced. A leak differs from an announcement in one way that decides everything else: nobody chose the moment. A planned sequence turns into a reactive one, audiences learn from outside, and the company faces a question it has not yet settled how to answer.

The mechanism is not a finance mechanism at all. Start somewhere ordinary. A household decides to sell the flat it has lived in for nineteen years. There is an order in which people are meant to hear it. The two children hear first, over a weekend, having grown up in those rooms. Then the tenant on the top floor, who has to be given time. Then the neighbours, who will want to know who is moving in. The order is a considered one, and every step of it exists to stop somebody hearing about their own future from a stranger.

Then a cousin mentions it at a wedding. By Monday the tenant has heard from the neighbour. Nothing about the sale has changed. The price is the same, the buyer is the same, the timing is the same. The household no longer holds the sequence. Three conversations now have to happen at once, badly, in an order the household did not pick, starting with the one person it most wanted to handle carefully. The wedding remark is a leak, complete in every part, and moving the same event onto a listed company and a Rs 1,140 crore purchase adds scale and paperwork without adding a single new idea.

Harivansh Packaging Limited, an invented maker of rigid and flexible packaging listed on both Indian exchanges, is buying all of Sundarban Polymers Private Limited, an unlisted maker of flexible packaging films selling into a customer list that overlaps the buyer's own. Devyani Kulkarni is the chief financial officer of Harivansh Packaging Limited and Ashwin Rege leads the transaction team. Twenty two weeks ran from the term sheet to completion on this transaction, of which the last nine were the conditions periodThe stretch between signing and completion, during which the things the agreement requires before completion can happen are obtained or satisfied., so signing fell at week thirteen of the twenty two. The twenty two weeks are this transaction's own elapsed weeks. Elapsed time varies enormously from one transaction to the next, so no standard length exists to compare them against.

Two more figures travel with the case and both need their label. Harivansh Packaging Limited carries an illustrative share price of Rs 300/- and an illustrative market capitalisation of Rs 5,400 crore, each of which belongs to a stated as-of date rather than to any live market, and the purchase is funded with Rs 140 crore of the acquirer's own cash plus Rs 1,000 crore of new borrowing at a contracted rate of 9.0 per cent. A price reaction is a movement rather than a level, and no movement was ever measured on that share price. Why the gap cannot honestly be filled is worked out below.

What is a leak, and how is it different from an announcement?

An announcement and a leak can carry identical content and still be opposite events. Take the two apart properly. An announcement is the company putting information about itself into the market at a moment it selected, in words it drafted, through a channel it chose, with everybody who needed advance warning already warned. A leak is the same information arriving in the market at a moment somebody else selected, or at no selected moment at all, in words nobody at the company wrote, through a channel nobody at the company picked, with the warning list only partly worked through.

So the difference is not accuracy and it is not content. The difference is authorship of the moment. Asking who chose when this became public separates the two events cleanly, every time, without needing to know anything else about either.

The reflex reading of a leak is the wrong one. When a story turns out to be substantially right, the instinct inside a company is relief: the facts are correct, so there is nothing to correct, so the damage is limited. What the company lost was never the accuracy but the choice of when to speak and in what form. An accurate leak is still a leak. A story that is right in every particular still arrives before the employees were told, still arrives before the counterparties were approached, and still arrives on a morning when the person who was supposed to speak for the company is on a train.

There is a second thing a leak takes that is easy to miss. An announcement is complete. An announcement says what is happening, on what terms, with what conditions, and says so in one place, so a reader who wants the whole of it can get the whole of it. A leak is a fragment by construction. Somebody knew a piece, and that piece is what travelled. So the market receives a partial account and the company is left holding the complete one. No position is less comfortable: the company knows the story is incomplete, knows the incompleteness is misleading, and has not yet decided whether it is able to say so.

Two terms do the rest of the work. The controlled groupThe set of people who are meant to hold information about a live transaction, each added deliberately and recorded, rather than anybody who happens to find out. is the set of people who are meant to have the information, added one at a time and on purpose. A leak is information crossing the edge of that set without anybody deciding it should. Which people belong inside the group in the first place, and what handling the information carries once they are in it, was settled earlier in this sequence and is used here rather than rebuilt.

One more distinction, and it is the one that decides how the first hour is spent. A leak is not the same as a rumour. In a newspaper the two are often the same sentence. A rumour is a story circulating in the market. The story may have come from inside the transaction, from an inference, from a competitor, or from nowhere at all. A leak is a story that came from inside. From where a reader sits the two look alike. From where Devyani Kulkarni sits the two are different problems. A leak tells her that the group holding the information has an opening in it. A rumour does not.

Where do leaks actually come from?

The unhelpful answer is carelessness, and it is unhelpful because it points at nothing anybody can change. Somebody was careless, so somebody should be less careless. Nobody has ever been able to act on that. The useful answer names three routes. Each of the three is addressed by a different control, and one of them is addressed by no control at all.

The first route is people. Somebody who holds the information tells somebody who does not. The people route is the one everybody imagines, and it is real, but the picture of it is usually wrong. The mental image is of a person selling information. The ordinary version is far duller: a spouse asked why the last four weekends were worked, a colleague brought into a call because a spreadsheet needed checking and never told what the spreadsheet was for, a proud sentence to a friend about the size of the thing being worked on. The information does not travel because somebody was malicious. The information travels because human beings talk about their lives, and a transaction is somebody's life for four months.

The second route is documents. A file travels further than the person who made it intended. The file prints on a shared printer and is collected an hour later by the wrong person. A copy goes to a personal address so it can be read on a phone at night. Another sits on a device that goes home. The file name says exactly what the file contains, so anybody who sees the name learns most of it without opening the file. A document leaks differently from a person: a conversation ends and a file does not, so a file keeps leaking.

The third route matters most. Nobody plans for it, and no one has to say anything at all for it to work. Call it observable behaviourThe visible pattern of what people do, as distinct from what they say. Arrivals, absences, meetings and site visits are all observable without a word being spoken.. The same set of unfamiliar people arrives at the same office for the fourth Tuesday in a row. A conference room stays booked for a month under a name nobody recognises. A group walks the length of a plant that receives no visitors, taking photographs, and nobody on the floor can place them. Two senior people miss a scheduled town hall with no explanation. A finance team that files its numbers on the eighth of every month suddenly needs a fortnight of extra reconciliations.

Every one of those is information. Not a sentence about the transaction, but the shape of a transaction, visible from outside to anybody who cares to notice, and there are people whose work is precisely to notice. A supplier's driver, a competitor's sales manager, a journalist who has covered the sector for eleven years and knows what an adviser's team looks like when it walks into a lobby.

A confidentiality undertaking reaches the first two routes and does not touch the third at all. A signed undertaking binds a person who might speak and governs a file that might travel. An undertaking has nothing whatever to say about six people arriving at a factory gate on a Wednesday. No paper stops arrivals. The only control available against the third route is the design of the process itself: where meetings happen, how visits are scheduled, what a room booking is called, whether a site visit is folded into something that would have happened anyway. Process design is a decision made in week one, not a legal decision made at any point.

What happens in the first hour?

Almost nothing that feels like doing something, and that is the whole discipline. The first hour is for establishing what is actually out there. The first hour is not for responding, not for drafting, not for calling the board, and above all not for correcting the parts that are wrong.

Three things get established, in this order. First, what has been said, in the actual words used rather than in a summary of them. Second, in what form it was said, meaning whether this is a precise account naming both parties and a price or a loose paragraph about consolidation in the packaging sector. Third, where the story appears to have come from. Only the third says anything about the group holding the information.

The response to a precise account and the response to a vague one are not the same response, and choosing between them requires knowing which of the two has appeared. So the order and the patience both matter. Answering the wrong version of a story is the single most efficient way to confirm the right one. If a journalist writes a loose sentence about consolidation in the sector and a company answers with a specific denial about a specific target, the company has just told that journalist which target to ask about next, and it has done so in writing.

The next step is genuinely counter-intuitive. Sit with it for a moment. The instinct when a story is partly wrong is to correct the wrong part, and correcting the wrong part is precision. Precision, offered to somebody who did not have it, is a disclosure. The disclosure is made by accident, in an unplanned minute, by whoever happened to be holding the phone. No worse combination of circumstances exists for releasing information about a live transaction.

The everyday version is a street vendor being asked by a rival where he buys his oil. If he says nothing, the rival learns nothing. If he says he does not buy it from the wholesaler on the corner, the rival now has a list one shorter and a direction to walk in. Denial of a specific is a statement about the general, and everybody hearing it knows that.

The first hour has one more job, and it is the quietest one. Somebody has to write down the time the story appeared, what it said, and who saw it first. Not for a report, and not for anybody's file, but because a week later, when people are trying to work out how far the information had travelled before it surfaced, the only honest account of the morning will be the one written during it. Memory reorders events under pressure, reliably and in one direction. The direction is towards whichever explanation everybody has since settled on.

What are the response options, and who chooses between them?

There are three, and they are worth naming flatly before anything is said about which is right.

The first is to say nothing, under a standing policyA position a company adopts in advance and applies every time, so that its answer on any single occasion carries no information about that occasion.. The important word is standing. A company that has always declined to discuss market speculation, on every occasion, in the same words, gives away nothing by declining again. The decline is uninformative. A company that usually answers questions and declines this one has said something quite loud. The value of a standing policy is built up over years of dull consistency and it is spent in a single inconsistent morning.

The second is to issue a prepared holding statementA short statement drafted and approved before it is needed, saying only what the company is willing and able to confirm, and issued unchanged when the moment comes.. The communication plan built this option in advance, and its whole merit is that it was written by people who were calm, checked by people who were careful, and approved before anybody was under pressure. A holding statement confirms only what can be confirmed and goes no further. The plan that produced it was settled earlier in this sequence.

The third is to announce. Announcing takes the announcement scheduled for a chosen future moment and brings it forward into an unchosen present one. Announcing ends the uncertainty. Announcing also ends every option that depended on the transaction not being public, and it is the only one of the three that cannot be reversed.

The choice is constrained rather than free, and the constraint is not a matter of judgement. The Securities and Exchange Board of India (SEBI), at sebi.gov.in, sets what a listed company must do when information about it reaches the market, when it must do it, and what may not be done with information about a live transaction by anybody holding it. Which of the three options is available to Harivansh Packaging Limited on any particular morning is a question for that authority. All three appear often enough to be worth recognising on sight.

The second point is about the person rather than the option. The choice must be made by the person the plan named in advance, not by whoever the first call happens to reach. Naming the person sounds procedural and is the most load-bearing rule in the plan. A leak arrives at a time nobody picked. So it arrives through whichever door happens to be open: a regional sales office, a plant reception, a personal mobile number a journalist kept from a conference three years ago. Every one of those doors is answered by somebody real, helpful, and entirely without authority to speak for the company on a transaction they may not know exists.

So the plan has to do two things at once. The plan has to name one person who holds the response. On this transaction that person is Devyani Kulkarni, chief financial officer of Harivansh Packaging Limited, working with Ashwin Rege, who leads the transaction team. And the plan has to give everybody else in both businesses one line they are authorised to say. The line is not a comment on the story but simply the name and number of the person to call. A line that short survives being repeated by somebody who has been ambushed at a factory gate. Anything longer does not.

What does a leak do to the process itself?

Most treatments skip this part. A leak's effect on the process is less dramatic than the morning of the story and costs a great deal more. A leak does three things to the transaction as a body of work, and it is worth taking them one at a time before noticing what they have in common.

The first is compressionThe shortening of a timetable under external pressure, so that steps designed to take a certain time are done in less of it.. Once a transaction is public, every day it stays unresolved is a day of open uncertainty, and open uncertainty is expensive to hold. Employees at both businesses are asking their managers questions the managers cannot answer. Customers are asking their account handlers whether supply arrangements are about to change. So the timetable shortens, not because anybody decided the work needed less time, but because the pressure to end the uncertainty exceeds the pressure to finish the work properly. Confirmatory diligence gets four days instead of a fortnight. A consent conversation that wanted three meetings gets one.

The second is the loss of optionalityThe ability to choose later. In a transaction it means the ability to stop, to change the terms, or to walk away, without that choice being observed.. Before a transaction is public, either side can stop. Stopping costs money, stopping costs time, and stopping is a private decision. Afterwards, stopping is a public event that a market watches, reads and interprets. Nothing about the arithmetic of walking away changed. The Rs 1,320 crore enterprise value did not change, the bridge from it to the Rs 1,140 crore equity value did not change, and the Rs 1,000 crore of new borrowing did not change. The buyer must now weigh the transaction against the cost of being seen to abandon it, and those are not the same weighing.

The third is the movement in price expectations. The third effect is the most direct. A seller's negotiating position is made almost entirely of alternatives: other buyers, or the credible option of not selling at all. Before the leak, the sellers of Sundarban Polymers Private Limited knew that Harivansh Packaging Limited was interested and nobody else did. After the leak, everybody who might have been interested knows that Sundarban Polymers is for sale and that somebody is already at the table. Knowing both facts is exactly what produces a second approach. The sellers' alternatives improved without the sellers doing anything.

Now notice what the three have in common. Every one of them runs against the party that was managing the process most carefully. The party that built a timetable loses the timetable. The party that preserved the ability to walk away loses the ability to walk away quietly. The party that negotiated a price against a known set of alternatives finds the alternatives have widened. A leak is not a symmetrical accident that lands on both sides alike. A leak is a transfer, and it moves value away from whoever had most invested in an orderly sequence.

The household version makes the asymmetry obvious. Take it again. If the flat's sellers had one interested buyer and the whole street learns the flat is going, the sellers gain: three more people call. The buyer found the flat first, negotiated quietly and thought the matter settled, and now finds himself bidding against people who would never have known. Nothing about the flat changed. The number of people who know changed.

Does it matter where in the sequence the leak lands?

Where the leak lands matters enormously, and not in the direction most people expect. Commit to an answer below before reading the reason. Being wrong out loud is what makes the reason stick.

Which audiences were not due to be told yet, and what happens to them?

Put the leak on this transaction's own timeline and be exact about the base. Twenty two weeks ran from the term sheet to completion. The last nine of those were the conditions period. So signing fell at the end of week thirteen, twenty two less nine. The conditions period covers weeks fourteen to twenty two inclusive, nine weeks in all. Confirmatory diligence and documentation fill the thirteen weeks before signing. Take the hardest case, a leak in week thirteen, on the morning the agreement binds.

Count them properly. Eight audiences sit on this transaction, and each one carries the week it was due to be told. The buyer's transaction team and the seller's transaction team knew from the approach, so they sit at week one. The board and governance of Harivansh Packaging Limited that had to authorise the term sheet sit at week two. The lenders sounded on the Rs 1,000 crore of new borrowing sit at week seven, when documentation begins and the funding has to be papered. The four inside the process have been told.

The other four have not. The first consent-holding counterpartyThe other side of a contract. Here it means a business whose own agreement with the target is affected by the transaction. was due to be approached in week fourteen and the second in week sixteen, both inside the conditions period. Their consents are conditions to completion. The employees of both businesses were due to hear at the announcement in week fourteen, the day after signing. The shared customers were due to hear in week eighteen, once the employee conversations had happened. So a leak in the signing week finds exactly half the audiences told and exactly half learning about their own future from a news report.

Now work the consequence that costs the most. The employees are not it, painful as that conversation is. The two counterparties are. Their consents are conditions to completion, so the transaction cannot complete without them. The condition was always true, and until the leak the counterparties did not know it. A planned conversation would have reached each of them with the transaction explained, the effect on their own contract set out, and a clear request. Instead each of them opens a news report and discovers, in the same paragraph, that its contract is about to change hands and that the transaction cannot close unless it agrees.

A counterparty that discovers its own leverage from a news report holds a stronger position than one approached in a planned conversation, and nothing about the transaction's economics changed to give it that. The Rs 1,320 crore enterprise value is the same. The Rs 1,140 crore equity value that actually reaches the sellers is the same. The Rs 1,000 crore of new borrowing at its contracted 9.0 per cent is the same. Every number in the transaction is where it was on the day before, and one party to it has become materially harder to satisfy, purely because of what it now knows about the position it was always in.

Leverage works the same way in any negotiation, not only in this one. Leverage is not made of rights alone. Leverage is made of rights the holder knows about and can see the value of. A third-party consentPermission from somebody outside the transaction whose own contract is affected by it, without which the transaction cannot complete. is worth one thing when it is requested as part of an orderly process nobody outside has heard about, and quite another when it is requested from a party that has just read what it is holding.

The audiences carry one more consequence, and it is the rule from the communication plan reasserting itself in a harder form. The plan's whole purpose was ordering: a sequence in which no audience is left to learn its own future from somebody outside, and in which each one is told by a person able to answer what it then asks. The moment a leak happens that ordering is already broken, and the recovery is simultaneity rather than a stubborn return to the original order.

Simultaneity over order is a genuine reversal, and it catches people out. The plan was carefully made, so the instinct after a leak is to stick to it. But the plan's order only ever protected people from hearing late from the wrong source, and once the story is public everybody is hearing from the wrong source at the same time. Continuing to work down a list means the sixth audience on it spends two days knowing that the company is talking to other people and not to them. The wait is a second injury, delivered deliberately, on top of the first one delivered by accident.

Work the control to the two ends and the finding sharpens. Move it to week one and six of the eight audiences learn from outside. Six sounds like the worst case and is not. In week one there is no agreement, nothing is settled, and the honest position is that a transaction may or may not happen. Move it to week eighteen and none of the eight learn from outside. Zero sounds like the best case and is not. By week eighteen the transaction is signed, the conditions are running, and the company is public with a commitment it can no longer quietly step away from.

And notice the flat stretch in the middle. No audience was scheduled to be reached between week seven and week thirteen, so a leak anywhere in that stretch produces the identical count, four told and four outside. The count of audiences is the same across those seven weeks and the situation is not remotely the same, so the milestone matters as much as the count. A leak in week seven finds a documentation draft and no binding agreement. A leak in week thirteen finds a binding agreement and nine weeks of conditions still to run.

What is done afterwards, and what can honestly be established?

Some time after the noise, there is a reviewA structured look back after an event, comparing what was recorded with what happened, in order to change something rather than to attribute blame.. Its method is simple to describe. Take the list of everybody who was recorded as holding the information, and set it against what actually escaped. Three things matter: how precise the story was, what details it had and what it did not, and which parts of the transaction a person would have had to be near to know them.

The comparison does real work. A story that carried the enterprise value but not the funding structure came from somebody who saw one set of papers and not another. A story that named Sundarban Polymers Private Limited but got the buyer wrong probably did not come from inside the buyer at all. Each detail present and each detail absent narrows the field, sometimes considerably.

The comparison almost never closes the field, and a review that pretends otherwise has done more damage than the leak it was investigating. The narrowing usually stops at a group rather than a person, for reasons that are structural rather than a failing of the review. Several people saw the same papers. The third route, observable behaviour, leaves no record at all and cannot be traced to a name by construction. And the information may have travelled through two or three hands before it surfaced, so the person nearest the story may be nowhere near the opening.

A review that ends without a name feels like a review that failed, and somebody senior asked for an answer. So the temptation is to produce a name anyway. The cost of yielding to it is severe and it is paid twice. A person is marked on the strength of proximity rather than evidence. And every future review inside that company becomes an exercise everybody is frightened of. The next one gets vaguer answers, from fewer people, later.

The output that repays the work is a changed control, not a name. Something concrete and small: the list of people holding the information stops being added to by whoever needs a spreadsheet checked and starts requiring one named approver. Site visits get folded into an audit that was happening anyway. Documents that leave the room carry a per-recipient mark. Meetings move to a location that explains itself. Each of those is a specific answer to a specific route, and each one can be pointed at a year later as the thing this cost bought.

Why can a leak be given no effect on the share price?

Because there is no such record, and the honest thing to do with an absent figure is to name the absence rather than to fill it.

The figures that do exist are worth stating precisely. Harivansh Packaging Limited carries an illustrative share price of Rs 300/- on 18.00 crore shares. Multiplying the two gives the illustrative market capitalisation of Rs 5,400 crore, and both figures belong to a stated as-of date that travels with them. Both figures describe a state. A price reaction is a different animal entirely: a movement, measured over a window, against some expectation of what the price would otherwise have done. No movement, no window and no expectation was ever recorded on this transaction, so no reaction can be computed from it.

Inventing a movement would be trivial. A sentence saying the Rs 300/- share price fell 4 per cent on the morning of the story would read perfectly well, would be believed, and would be entirely made up. A reader given a number carries the number away and leaves the mechanism behind.

Of the two, only one is likely to be still held in six months: that a leak in the signing week reaches the counterparties whose consents the transaction needs before anybody has spoken to them, rather than that some invented company's invented share moved by some invented percentage. The first travels to any transaction a reader ever looks at. The second was never true of anything, so it travels nowhere.

There is a second reason and it is about what a price movement would even mean. A market seeing a story about an acquisition is reading several things at once: whether the transaction is likely to happen, what it might cost, how it might be funded, and what the mere existence of a leak says about how tightly the buyer runs its processes. Attributing a single number to one of those is a claim about a market's reasoning, and no invented case can support a claim about how a real market reasons.

How does a lender, an analyst or a counterparty actually read a leak?

Four readers pick the same story up and reach immediately for four different first questions, and not one of those four is what the article itself was written to answer.

A lender who has been sounded on the Rs 1,000 crore of new borrowing reads a leak as a change in the probability that its own commitment gets drawn, and as a change in how public its position has become. The lender had been asked, under confidentiality, whether it would fund something. Now the something is in the newspaper, and the lender's credit process has to reckon with the possibility that the transaction accelerates, that the terms move, or that it collapses noisily. Its first question is not about the target at all. The question is whether the borrower's process is still holding. A borrower whose process leaked in week thirteen is a borrower whose conditions might also slip.

An analyst covering Harivansh Packaging Limited reads it as an incomplete set of facts arriving out of order, and the first thing a careful one does is write down what is not in the story. Is there a price? Is there a structure? Is there anything about funding? The absence of the funding is usually the most informative gap. A Rs 1,000 crore borrowing against an existing Rs 740 crore of borrowings is a different proposition from an all-cash purchase, and a story that does not carry the funding has not been sourced from anywhere near the funding papers. The analyst is reading the story for what it says about its own origin as much as for its content.

An investor already holding the shares reads it as a question about process quality rather than about the Rs 1,140 crore purchase. Whether this purchase is a good idea is not settled by a news report and will not be settled by one. A leak does tell a holder something about how the company runs itself, and that observation about management is durable in a way that any single transaction is not.

A counterparty whose consent is a condition reads it as the arrival of information about its own position, the reading already worked through above. The counterparty learns that its contract is in play and that the transaction cannot complete without it. Everything the counterparty does next follows from that knowledge, and none of it follows from the transaction's economics. The economics are exactly where they were.

The household version holds for all four. If the neighbours learn that a flat is being sold before the tenant does, the tenant does not primarily conclude something about the flat. He concludes something about the household selling it, and he adjusts how much warning he expects to get about anything else. The adjustment outlasts the sale.