Transaction Risk and Integration Risk: Where the Line Falls

Transaction risk is anything that can move the price or stop the purchase before money changes hands, and it is settled by documents. Integration risk is anything that can go wrong once the buyer holds the business, and it is settled by operations. The line between the two is the completion date. An earn-out is the one term that sits on both sides of it.

What is transaction risk, and what settles it?

Consider a small provision shop bought from the person who has been running it for twenty years. The price is agreed in March on the strength of what the buyer saw: the stock on the shelves, the money in the till, the amount the shop still owes its suppliers. Handover comes in June. In those three months the stock can be run down, the till can be emptied, a supplier can decide the arrangement stops when the shop changes hands. None of that is dishonest and none of it is dramatic. The price was agreed for a picture, and pictures move.

Everything in that gap has one thing in common. Every single item can be dealt with by writing something down before the money changes hands. The agreement can say that the price will be adjusted for whatever stock is actually on the shelves on handover day. The agreement can require the supplier to confirm in writing that it will carry on. The agreement can say that if the supplier refuses, the buyer does not have to complete at all.

Transaction risk is the whole set of things that can move the price or stop the purchase between the day terms are agreed and the day money changes hands, and every one of them is answered by drafting rather than by doing. The shared answer is not a coincidence and not a definition chosen for tidiness. Being answerable by drafting is the property that makes the category worth having.

Put the shop away and look at the purchase this sequence has been working through. Harivansh Packaging Limited has agreed to buy the whole of Sundarban Polymers Private Limited. Enterprise value is Rs 1,320 crore. Take off the Rs 180 crore of net debt that Sundarban Polymers carries. The equity value, the amount that actually reaches the sellers, is Rs 1,140 crore. Between signing and completionThe point at which the shares actually transfer and the money is paid. Signing and completion are two different dates, and a great deal happens in between., three sorts of thing can happen to that Rs 1,140 crore.

The balance sheet can move. Working capital and net debt are living numbers; they are different on a Tuesday from what they were on a Friday. The agreement handles this with a peg and an adjustment, and completion accountsA set of accounts drawn up as at the completion date, used to measure the items the agreement said would be trued up, such as working capital and net debt. measure what actually happened. A peg and an adjustment are drafting.

A condition to completionSomething that has to happen, or be obtained, before either side is required to complete. If it is not satisfied, the parties are not obliged to go ahead. can fail. An approval does not arrive; a consent is refused; a party does not deliver what it promised to deliver. The agreement handles this by saying what happens if the condition is not met. A stated consequence for a failed condition is drafting too.

Something can be found late. Due diligence turns over a stone in the last fortnight and there is a liability under it. The agreement handles this with a warrantyA statement of fact made in the agreement by one party to the other. If the statement turns out to be untrue, the party that relied on it can bring a claim., an indemnity, a price reduction or a withdrawal. Drafting again.

Notice what the three have in common. The three are about very different subjects: an accounting balance, a regulatory approval, a legal exposure. Subject is not what relates them. The relation runs through the response instead: each of the three is answered by a sentence in a document, agreed by two parties who are both still at the table, before either of them is committed beyond withdrawal. Devyani Kulkarni, the chief financial officer of Harivansh Packaging Limited, does not manage transaction risk. She negotiates it.

What is integration risk, and what settles it?

Back to the provision shop, now past handover day. The buyer stands behind the counter. The stock has been counted, the price has been adjusted, the supplier has confirmed in writing that it will keep delivering. Everything that was written down has been done. And on the first Monday the regulars come in, see a face they do not recognise, buy what they came for and then start buying two of the six things they used to buy here and four of them from the shop two streets away.

No amount of drafting reaches that. There is no clause in any document anywhere that makes a customer come back. There is no sentence that could have been written in March to keep the boy who knew where everything was on the shelves from taking a job elsewhere in July. There is no paragraph that makes the new way of stacking the shelves and the previous holder's way of stacking them become one way without a period of nobody being able to find anything.

Integration risk is the whole set of things that can stop the business performing the way the price assumed, once the buyer holds it, and not one of them is answered by drafting. They are answered by management: by turning up, deciding, hiring, explaining, choosing what to keep and what to change, and living with the results for years.

In this purchase, integration risk is the whole of what happens after Harivansh Packaging holds Sundarban Polymers. Sundarban Polymers sells to some of the same customers as Harivansh Packaging, and that overlap is the reason the purchase exists at all. The overlap is also the first place things can go wrong. A customer that was buying from both of them now has one supplier where it had two, and may decide it would rather have two. The decision is made in somebody else's meeting room, on a date nobody controls, for reasons nobody in this purchase will be told.

Then there are the people. The people who have been running Sundarban Polymers know which machine runs hot, which customer will accept a late delivery and which one will not, and which order is worth taking at the quoted price. Some of those people will stay and some will not. A buyer can offer money to encourage them to stay, and money helps, and money is not the same thing as a mechanism.

Then there are the two ways of working. Two production footprints have to be scheduled together or kept apart on purpose. Two sets of quality practice, two ways of pricing a small order, two ways of deciding when to say no. Merging two ways of working is slow, and the slowness itself has a cost. While the merge is going on, both businesses are being run by people who are also in meetings about being run.

Here is the property that matters, and it is the mirror image of the one on the other side. Every integration risk that can be named is settled by somebody doing something for a long time. A document that promises to address integration risk usually cannot deliver the promise. A promise about an outcome is only worth what it can be enforced as, and nobody can be made to deliver an outcome they do not control.

Where exactly does the line between them fall?

Not in a grey zone, and not over a period. The line is the completion dateThe single named day on which the shares transfer and the consideration is paid. Before it, neither side holds what the other has agreed to give., and a completion date is a specific day with a specific date on it, written into the agreement and later recorded in the filings.

The completion date is the line, it is one day rather than a stretch of time, and the same event falling either side of it lands on a different party and gets a different response. The last clause is the useful part, so take it slowly.

Suppose a large customer of Sundarban Polymers gives notice that it is moving its business elsewhere. If that notice arrives three weeks before completion, it is a transaction risk. The notice arrives while both parties are still at the table and neither is committed. The price can be renegotiated, a specific indemnity can be asked for, the buyer may in some drafting be able to walk away. The loss, if any, is shared by negotiation, and whatever is agreed is written into the document.

If exactly the same notice, from exactly the same customer, for exactly the same reason, arrives three weeks after completion, it is an integration risk. The money has gone. The shares have transferred. The sellers are no longer parties to anything except whatever survives in the agreement. There is nobody to renegotiate with. The buyer bears it, and the response is not a clause but a plan: hold the volume elsewhere, cut the cost, find another customer.

Same event, same magnitude, same cause. Different side of one day, and therefore a different party carrying it and a different kind of response. A different party and a different response are what make the completion date worth naming as a line. Most people leave the line implied and carry it around that way.

Which risks look like one and are actually the other?

A reader who has followed everything so far gets caught here. The trap is that most people sort risks by what they are about. Anything that mentions a customer feels like an operating matter. Anything that mentions a balance sheet feels like a transaction matter. Both instincts are wrong often enough to be expensive.

Take the first crossover case. Sundarban Polymers has a supply contract with a term saying the counterparty may terminate if there is a change of controlA clause allowing a party to end or renegotiate a contract when the shares of the other party pass into new hands. The trigger is the transfer itself, not anything the business does. of Sundarban Polymers. Read by subject, this is about a supplier, and suppliers are operations, so it looks like integration risk. Read by settlement, it is nothing of the sort. The clause is triggered by the transfer of the shares. The transfer happens on the completion date. The clause can be dealt with before that day by going to the counterparty and obtaining a consent or waiverA written agreement from the counterparty that it will not use its right to terminate, given before the event that would trigger that right., and if the consent is refused, that fact can itself be made a condition to completion. Drafting, consent, condition. A change of control clause is a transaction risk wearing an operating costume.

Now take the second crossover case, and read it the same two ways. Working capital at Sundarban Polymers has been managed downwards in the weeks before completion. Receivables have been chased hard, payments to suppliers have been left a little longer, stock has not been replaced quite as promptly. Read by subject, the position is about how the business is being run day to day. Day-to-day running sounds like operations. Read by settlement, it arrives as a number in the completion accounts and changes the price. An operating fact, presented as a price adjustment, and settled by a mechanism agreed in advance.

The sorting test is not what a risk is about, it is what settles it, and the two crossover cases show that the subject matter of a risk is an unreliable guide while the settlement mechanism is a reliable one. Anything answered by a sentence in a document sits before the line. Anything answered by somebody running the business sits after it.

The same trap is easy to feel in an everyday version. A tenancy on a small workshop is being taken over. The landlord has a clause letting him end the tenancy if the workshop changes hands. The landlord clause sounds like an operating problem and is not. The clause is settled in advance by asking the landlord to sign a letter, and if he will not sign, the incoming tenant does not sign either. Whether the neighbours keep sending work after the handover is a different sort of thing entirely, and no letter from anybody settles it.

Which single term sits on both sides of the line?

Almost every term in an agreement sits cleanly on one side. The adjustment mechanism is before. The conditions are before. The warranties are before as well. A warranty states a state of affairs at or before completion, whatever date a claim on it is brought. The plan for the first year is after. There is one exception in this purchase, and the exception is what makes the completion date a line that a single term can straddle.

The earn-outA part of the purchase price that is not paid at completion but becomes payable later, and only if the business reaches a stated result within a stated period. is that exception. The agreement provides that a further Rs 60 crore is payable if Sundarban Polymers reaches earnings before interest, tax, depreciation and amortisation (EBITDA) of Rs 145 crore in the first year after completion. Sundarban Polymers earned Rs 132 crore, so the thresholdThe stated level a result must reach before the conditional payment becomes due. Below it, nothing is paid; at or above it, the stated amount is. asks for Rs 13 crore more, which is 9.8 per cent above where it started.

The amount and the threshold are fixed by a document signed before completion, and whether the amount is ever paid is decided by how the business is run after completion, so the same single term has one part on each side of the line. Nothing else in this purchase has that shape.

The two halves behave completely differently, so look at each one separately. The half that sits before the line is as bounded as anything else on that side. Rs 60 crore is a number. Rs 145 crore is a number. The period is one year. The definition of what counts as EBITDA for this purpose is a set of words, and those words are agreed while both parties are still at the table. All of it is negotiated, drafted and closed on or before the completion date, exactly like the working capital peg.

The half that sits after the line behaves like every other integration risk in this guide. Whether Sundarban Polymers reaches Rs 145 crore depends on whether the customers stay, whether the people stay, whether the two production footprints work together and whether the buyer moves anything into or out of the business. The agreement settles none of that, and nothing in the list could be settled by an agreement.

Earn-outs attract sentiment in both directions. The consequence allows neither. An earn-out converts part of the integration risk into a priced transaction term. The conversion is a transfer, not a trick played on the sellers and not a solution to the buyer's problem. The risk that the business does not perform has not gone anywhere. Somebody still has to run the business, and the result will be whatever the result is. The change is in who carries the consequence of the result. Before the earn-out, the buyer had paid a fixed amount and carried the whole of the outcome. After it, Rs 60 crore of the amount rides on the outcome, so the sellers carry that much of it.

The everyday version is a house sale where part of the price is held back until the roof gets through one monsoon. The holdback does not make the roof better. The holdback does not make the rain lighter. The holdback decides who pays if the roof leaks, and it does so in advance, in writing, for a stated amount.

What did the transaction risk actually cost, and what rides on integration?

Sorting is only worth doing if the two columns look different when the figures go in. Every figure below was settled earlier in this sequence and is carried in.

Start with the side that has numbers. The agreement set a normalised working capital of Rs 96 crore. The completion accounts showed Rs 108 crore, so the price adjusted up by Rs 12 crore. The agreement assumed net debt of Rs 180 crore. The completion accounts showed Rs 195 crore, so the price adjusted down by Rs 15 crore. Plus Rs 12 crore and minus Rs 15 crore is a net of minus Rs 3 crore, and the equity value paid moved from Rs 1,140 crore to Rs 1,137 crore.

Against the Rs 1,140 crore headline, that net movement is 0.26 per cent. Transaction risk landed, it was measured and it was settled on one day, bounded in advance by a mechanism both parties had agreed. Bounded in advance by an agreed mechanism is what defines the whole category. Nobody argued about whether working capital had moved; they read a number off a set of accounts.

The small net is the most misread figure in the comparison. The Rs 3 crore is small because two movements ran in opposite directions, not because either movement was small. Working capital alone moved the price by Rs 12 crore, or 1.05 per cent, and net debt alone moved it by Rs 15 crore, or 1.32 per cent. A buyer who checked only working capital would have been wrong by Rs 12 crore, and one who checked only net debt would have been wrong by Rs 15 crore. The smallness of the net is the reason both are computed separately and never the reason either is skipped.

Now the other column, and notice at once what is missing from it. Whether Sundarban Polymers keeps the customers it shares with Harivansh Packaging. Whether the two production footprints can be run together. Whether the people who have been running the business continue to run it. Not one of those has a figure anywhere in the agreement. The absence is not a gap in the record but the defining property of the column. A thing settled by operations has no number in a document, because a document cannot settle it.

There is, however, one figure that stands next to the integration column, and it is the straddling term. Rs 60 crore rides on the earn-out. Against the Rs 1,137 crore actually paid, that is 5.28 per cent.

A percentage without its base is not a reading, so set the two against each other with the bases named. The transaction risk that actually landed moved the price by Rs 3 crore, or 0.26 per cent of the Rs 1,140 crore headline it moved away from. The amount riding on integration is Rs 60 crore, or 5.28 per cent of the Rs 1,137 crore that was actually paid. Each figure is struck on the amount it belongs against, and the two are deliberately not forced onto one base. One measures a movement from the headline and the other measures an exposure against the settled price.

The second figure is roughly twenty times the first and carries no mechanism that bounds the outcome behind it. A bounded risk that certainly lands is not obviously smaller than an unbounded one that may never land at all, so the two figures size each other without ranking each other. Size and ranking are two different claims, and a reader who takes the size for a ranking has taken the wrong lesson.

Why does treating the two as one thing produce the wrong response?

Because the response is the only thing that distinguishes them, so a team that has collapsed the two categories has thrown away the only information the sort was carrying. The failure runs in both directions and looks quite different each way.

In the first direction, a team treats integration risk as something to draft against. The team asks for a warranty that the ten largest customers will still be buying in two years. The team asks for an undertaking that the senior people will stay. The team asks for a promise that the combined business will hold its margin. Every one of those requests is about a real risk, and every one of them is a promise about an outcome that the person giving it no longer controls. After completion the business is being run by the buyer. A clause written against an outcome nobody controls is not protection. The clause converts an operating problem into a dispute rather than into an answer, and a dispute settles years later at a cost neither party priced. Worse, the request itself buys false comfort: a team that has obtained the words often stops doing the work.

In the second direction, a team treats transaction risk as something to manage later. The team notices that working capital looks a little thin, decides it will keep an eye on it after taking over, and does not put a peg in the agreement. The team hears that a supply contract has a change of control clause, decides it will build a relationship with the counterparty once it is in charge, and does not ask for a consent. Both decisions feel practical and both are irreversible. Once the money has gone across, a transaction risk that was not settled has been paid for in full, and there is nobody left to renegotiate with. The mechanism that could have handled it, at almost no cost, existed only while the other party still wanted the transaction to happen.

Look at how ordinary both errors are. Nobody in either version was careless. In the first version the team was thorough. In the second it was pragmatic. The mistake in both is the same mistake: sorting by subject and effort rather than by settlement, and therefore aiming the response at the wrong side of one day.

How does a practitioner use the sort before negotiating?

By turning it into a list, at the start of the work rather than at the end, and by using it in three different jobs that read the same purchase from three different chairs.

The sort decides where the drafting hours go, so the transaction team runs it first. Ashwin Rege, who leads the transaction team at Harivansh Packaging Limited, takes the term sheet and puts every item into one of three groups. Before the line: pegs, adjustments, conditions, consents, warranties, the mechanics of completion. After the line: retention of customers, retention of people, the schedule for putting two production footprints together, anything about how the business will be run. Straddling: the earn-out, and in this purchase nothing else. The straddling group gets the definitional work first, and the after group gets a named person rather than a clause.

A lender reads the same list for a different reason. A lender to Harivansh Packaging Limited is funding Rs 1,000 crore of new borrowing and wants to know two things about the risks in front of it. First, how large the day one funding requirement can get: that is the transaction column, and because it is bounded by a mechanism, a lender can size it. On these figures the mechanism moved the requirement by Rs 3 crore net, though it could have moved it by more, and the mechanism itself is what makes the range knowable. Second, what future calls on cash the borrower has already committed to: the earn-out of Rs 60 crore is exactly that, a possible payment in the first year after completion that does not appear in any borrowing figure today.

An analyst reading the announcement uses the sort as a set of questions rather than as a conclusion. Only one of the two figures is what the sellers receive, so which of them is the equity value and which the enterprise value. Whether the price quoted is before or after the completion adjustments. Where there is an earn-out, which EBITDA the multiple was struck on. The last question is not pedantry. Maximum equity value here is Rs 1,137 crore plus Rs 60 crore, or Rs 1,197 crore, and with the Rs 195 crore of net debt measured at completion the maximum enterprise value is Rs 1,392 crore. Rs 1,392 crore is 10.55 times the Rs 132 crore Sundarban Polymers actually earned and 9.60 times the Rs 145 crore the payment is conditioned on. Both readings are true, they describe different things, and an announcement or a note quoting one of them without saying which EBITDA it used has not finished the sentence.

The sort identifies which risks take document time, which take management time, and which single one takes both, and it identifies them before any of the time has been spent.

Which kind of risk matters more?

The question has no general answer, and the reason is structural.

The two are not on a common scale. Transaction risk is measured as an amount, in one currency, on one day, against a mechanism that both parties agreed in advance. Integration risk is measured as performance, over years, against an expectation that only exists inside the price. Rs 3 crore can be set next to Rs 60 crore, as it has been above, and the ratio of twenty is a real ratio. Rs 3 crore cannot be set next to a customer relationship, and the Rs 60 crore earn-out is not the size of the integration risk. Rs 60 crore is the size of the part that was written into a document.

Any ranking of the two would depend on the particular business rather than on anything general, so sorting them and sizing them is as far as the general case reaches. A purchase of a business with one large customer and three key people is mostly integration risk whatever the adjustment clause says. A purchase where the balance sheet swings hard between quarters and half a dozen approvals are outstanding is heavy on transaction risk before anybody gets to the first Monday. The sort establishes which is which. The weighting is a fact about the particular business.