The Non-Compete: Stopping the Seller From Rebuilding

A non-compete is the seller's promise not to start or join a competing business, for a stated time, in a stated place, over a stated activity. The promise guards the part of the price that never sat on a balance sheet. Harivansh Packaging Limited paid Rs 820 crore more than the net worth of Sundarban Polymers Private Limited, an invented maker of flexible packaging films, and customers, know-how and reputation leave in the same cars as the people who carry them.

Three numbers from this purchase are already settled. Sundarban Polymers Private Limited earns Rs 132 crore of earnings before interest, tax, depreciation and amortisation (EBITDA) on Rs 880 crore of revenue, a margin of 15.0 per cent. Ten times was the multiple the two sides settled on, and ten times Rs 132 crore put the whole business at Rs 1,320 crore. Its lenders were owed Rs 180 crore net of the cash in the till, and that obligation stayed inside the company when the shares changed hands, so what the sellers were handed was Rs 1,140 crore. One kind of promise is the warrantyA statement in the agreement that something about the business was true on a stated day. Where it was not, the buyer has a claim for money. Set out in full under warranties and indemnities.. A warranty is a promise about what was already the case. The other kind of promise is the restriction, and the two behave nothing alike.

What is the seller actually promising, and when does the promise start?

Any transaction document, opened at the promises, carries two species of sentence sitting quite close together, doing entirely different jobs. One species looks backwards. The seller says that the machines are owned, that the tax filings were made, that the customer contracts are what they appear to be. Each of those sentences describes a state of the world on a particular day, and the buyer's remedy exists because the sentence turned out not to describe it.

The other species looks forwards. The seller says it will not do a thing. The seller will not set up a competing business. The seller will not join one. The seller will not, for a stretch of time the paper names, in a place the paper names, over an activity the paper names, be found on the other side of a customer's desk from the buyer.

A warranty is a promise about what was true, and a restriction is a promise about what the seller will not do next. A breach of one tells nothing at all about the other. The distinction sounds obvious written down. On a single summary sheet reduced out of two hundred printed sides, every promise starts to look like every other promise, and the distinction stops being obvious.

The completion line has a consequence for the order in which a document gets read. When a buying team finishes diligence, its attention is almost entirely behind the line. Every finding, every schedule, every argument about a disclosure is about the past. The restriction is the one clause in the money-carrying half of the document that is looking the other way, and it is usually agreed late, in a paragraph, by people who are tired.

Here is the everyday version. A household buys a coaching centre from the teacher who built it. The benches, the rented room, the whiteboard and the printed notes all transfer, and the seller can honestly promise that the rent is paid up and the notes are the seller's own. Every one of those promises is about the past and every one of them can be true. The parents, meanwhile, enrolled their children because of the teacher. If the teacher opens a new centre two streets away in April, nothing anybody promised about the past has become false, and the thing the buyer paid for has gone.

Why is the money at stake the part paid above net assets?

Take the two figures side by side and the argument for the clause makes itself. Sundarban Polymers Private Limited's own books carried net assetsEverything a business has, less everything it owes, as its own books record the two. On this target that came to Rs 320 crore. of Rs 320 crore. The Rs 320 crore is the machines, the buildings, the stock in the yard, the money owed by customers, less everything owed to everybody else. Net assets are the part of the business a valuer could walk around and point at.

The buyer paid Rs 1,140 crore. Subtract the Rs 320 crore and Rs 820 crore is left over, and that Rs 820 crore did not buy a single object. The Rs 820 crore bought the fact that a large food packaging customer picks up the phone to this business and not another one. The same Rs 820 crore bought the knowledge of which resin blend runs cleanly on which line at what speed. And it bought a reputation for delivering when the monsoon has closed a road. Because none of those three is owned in any sense a document can transfer, every one of them was bought indirectly. Customers, know-how and reputation sit with people, and people can walk.

71.93 per cent is the share of the headline price that bought nothing anybody could walk up to and touch. Struck instead on the Rs 1,137 crore that actually left the buyer's account, the same measure reads 71.86 per cent. Neither proportion is doing the work; the rupees are, and they carry no rounding at all. Rs 820 crore set beside Rs 817 crore are Rs 3 crore apart, and Rs 3 crore is precisely where the two completion adjustments landed once both of them had run.

The Rs 3 crore gap is where the second base comes from, and one sentence explains it. The working capital position at completionThe day the shares actually move and the money actually goes across, which is usually some weeks after both sides have signed. came in Rs 12 crore above the level the agreement had pegged, and that pushed the price up. Net debt came in Rs 15 crore above what the transaction had assumed, and that pulled the price down. Plus Rs 12 crore and minus Rs 15 crore leave minus Rs 3 crore, so Rs 1,140 crore became Rs 1,137 crore and Rs 820 crore became Rs 817 crore.

Both bases are honest and they answer slightly different questions, so every figure has to carry the base it was struck on. At one decimal the two proportions are indistinguishable: 71.9 per cent and 71.9 per cent. At two places, 0.07 points separates them. The 0.07 point gap is real, and an argument that leaned on it would be leaning on rounding rather than on rupees.

A bigger business will always produce a bigger amount, so the amount on its own does not travel between transactions. The proportion travels, so divide the amount by the price it came out of.

Read the proportion aloud and it stops being a statistic. A little over seven rupees in every ten, 7.19 of them on the headline base, went to something that cannot be fenced, insured, depreciated or photographed. The restriction is the only clause in the entire document addressed to that seven-tenths. Everything else in the promises half of the paper, every warranty and every schedule behind it, is aimed at the three-tenths that a balance sheet already describes.

Which three settings does the paper choose when it writes a restriction?

Every restriction, however it is worded, sets three things. How long it runs. Where it applies. Which activity it covers. The three settings are not opinions and they are not hidden; they are written into the sentence and a reader can find all three in under a minute.

Each of the three is worth a minute because each one is a boundary, and a boundary has an outside. A restriction can be perfectly clear on all three settings and still leave the thing the buyer paid for entirely outside it. A clause of that kind is the failure mode that reads as competence: nothing is vague, nothing is missing, and nothing is covered.

The activity setting, put against this transaction, shows how that happens. Sundarban Polymers Private Limited makes flexible packaging films. A restriction written over rigid packaging would be a real sentence, imposing a real obligation, on a business the buyer did not acquire. Nobody drafted it dishonestly. Harivansh Packaging Limited makes both rigid and flexible packaging, so somebody drafted the clause from a description of the buyer rather than from a description of what was bought, and the wrong half of that sentence found its way in.

The same test runs on place and on time. A place setting is checked against where the customers are, not against where the offices are. A time setting is checked against how long the relationships take to move; a customer contract that renews annually and a plant qualification that takes two years to repeat behave very differently. A clear boundary and a useful boundary are not the same thing, so the job on this clause is to find all three settings and then ask what each one puts outside itself.

Whether a restriction of any width would bind anybody at all is a legal question, settled by lawyers and by the courts rather than by arithmetic. The width that appears in any particular document was negotiated by people who took advice on it.

How is a non-solicit a different promise from a non-compete?

A non-compete addresses one route by which value leaves: the seller sets up shop again and takes the customers by trading against the buyer. There is a second route, and it does not involve the seller trading at all. The seller simply stops being the place the key personSomebody whose leaving would change what the business is able to do, because the relationships or the know-how sit with them rather than in a system. works, and those people go somewhere else, and what they carry goes with them.

A non-solicit is the separate promise about that second route. A non-solicit says the seller will not approach, hire or take away the people, and in some documents the customers as well. The non-solicit is a different sentence in a different place doing a different job, and a document can carry one without the other.

Put a household on it again. A small catering business is sold. The seller promises not to open another catering business in the same city, and keeps that promise faithfully, cooking nothing for anybody. Six months later the two cooks who actually made the food and the one person every wedding planner in town rings have all gone to work for a rival. Nobody competed. The buyer bought a kitchen and a set of relationships and now has the kitchen.

The finance version at Sundarban Polymers Private Limited scale is identical in shape and larger in rupees. Four senior people carrying the key account relationships, the plant floor know-how and the pricing history walk from the target to a competitor. The sellers started nothing, joined nothing and traded against nobody. The non-compete has been kept in every word, and the value it existed to protect has left the building anyway.

One more turn of the same screw. The non-solicit has settings of its own, and it names what it covers just as narrowly: in most documents, the hiring of people. A promise about hiring answers a question about hiring, and a customer who chooses to move without anybody being approached is a third event again. Read the second promise exactly as carefully as the first: it is not a catch-all, it is another boundary with its own outside.

Why does a broken restriction not reach the same money as a warranty claim?

The machinery that filters a warranty claim is already in place. A floor keeps the small items out of the counting entirely. A basket of Rs 11.37 crore is the pile the counted claims must together clear before any money moves. A capThe most a buyer could ever recover under the promises the document says it applies to, however large the loss turns out to be. Worked in full further along here. of Rs 227.40 crore stops the whole thing. Twenty per cent was the fraction agreed, and the Rs 1,137 crore adjusted base is what that fraction was taken out of. An escrow of Rs 113.70 crore sits with a neutral holder so that a successful claim has something behind it.

Every one of those four numbers was negotiated with warranties in mind. The four numbers exist because a buyer and a seller had to agree how much of the risk that a statement about the past turns out wrong would sit on each side. A restriction is not a statement about the past. A restriction is a separate promise doing a separate job, and whether the same four numbers apply to it is not a matter of arithmetic or of general principle.

The answer is written down, and it is not written the same way in every document, so find the sentence that says which promises the limits apply to. The whole instruction is that one sentence. In some documents the limitation clause is expressed to cover claims under the agreement generally, and that sweeps the restriction in. In others it is expressed to cover claims for breach of warranty, and that leaves the restriction outside it. Both are ordinary drafting and the difference between them is worth more than most of the schedules.

A reader who does not look this up will make one of two errors, and both are expensive. Assuming the cap applies tells a credit committee that the maximum loss on this paper is Rs 227.40 crore. Assuming it does not tells that committee the buyer has unlimited recourse on a clause the other side never agreed to leave uncapped. Neither statement is knowable from the arithmetic; both are knowable from one paragraph.

What happens when part of the price is written against the restriction?

Most price clauses state one amount for the shares and stop. Some do not. Some split the confirmatory diligenceThe checking a buyer does after the outline terms are agreed and before the document is signed, to see whether the business is what it was said to be. outcome into a figure for the shares and a figure stated against the restriction itself. The paper then carries a rupee number for the promise not to compete.

An allocation puts a number on the restriction inside the document, and that is the whole of what it does for a reader. An allocation does not make the promise wider, narrower, stronger or weaker. Nor does it add a remedy. The allocation adds information: two people who negotiated this deal wrote down what they thought the seller staying away was worth, and that figure can be read.

The information is genuinely useful. A restriction with a large allocation against it was argued over. A restriction with a token allocation was probably agreed in a paragraph at the end of a long night. Neither settles whether the clause works. Both show how seriously the parties treated it, and that is a different and still worthwhile signal.

Money described as paid for something is treated according to what it was paid for, so an allocation also raises a tax question, immediately and unavoidably. How that treatment works is settled by the Central Board of Direct Taxes rather than by the agreement, and a reader meeting a split price clause has a second question standing behind the first.

How large is the exposure a restriction is actually aimed at?

The two halves of the document set beside each other in rupees make the comparison that carries the whole argument, and it takes one line of arithmetic.

Rs 817 crore is the exposure a restriction is aimed at, measured on the Rs 1,137 crore that the sellers were actually paid once both completion adjustments had run. The ceiling on warranty claims is Rs 227.40 crore, or 20.0 per cent of that same adjusted base, and Rs 817 crore is 3.59 times as large. On the headline base the two figures are Rs 820 crore and the same Rs 227.40 crore, a ratio of 3.61 times. The ceiling itself does not move, having been struck on the adjusted base to begin with.

Read the dashed line for a moment. The dashed line is where the ceiling would cut the exposure if the ceiling reached this clause at all. Everything to the right of it, Rs 589.60 crore of it, is beyond any number in the limitation clause under any reading. The two clauses were written to answer different questions, so the largest exposure in this transaction is not filtered by the largest number in the limitation clause.

The comparison claims less than it may appear to. Nothing in the comparison says that a seller who rebuilds destroys Rs 817 crore of value; nobody can know that figure. The comparison says only that the money the restriction is pointed at is several times the money the limitation clause is pointed at, and that a reader who has satisfied themselves about the cap has satisfied themselves about the smaller of the two.

Nor does the comparison say anything about how much protection any particular restriction gives. The reading check takes over from the arithmetic at that point.

What does a reader check when handed a restriction clause?

Four checks, and the order matters more than the list. Duration is usually the first thing the sentence says, so a reader working down from the top of the clause puts duration first. Worked in the other order, the checks save the trouble on the clauses that were never going to matter.

A restriction that binds a selling company and not the individuals who ran it has been given by the wrong party. Think about who is actually in a position to rebuild. A private company that has just been sold is a shell holding a bank balance; it has no customers, no plant and no intention of doing anything ever again. The two people who built the customer relationships over fifteen years are not that company. If the clause names only the entity, the promise has been extracted from the party with no ability to break it.

Beside those four sits a fifth question that is not about the restriction at all: does a separate promise about hiring sit next to this one? A reader who runs all five in the order above has read the restriction properly in about four minutes, and four minutes is a fair price for the clause pointed at 71.86 per cent of what the buyer paid.

Who reads this clause, and what does each of them want from it?

The same paragraph is read by four different people for four different reasons, and knowing which of the four is reading stops the wrong question being asked of it.

A lender to the buyer is sizing a facility against the buyer's cash flows, and after this purchase Harivansh Packaging Limited carries a great deal more borrowing than it did. A lender's interest in the restriction is narrow and specific: the business it lent against is partly a set of relationships, and it wants to know whether those relationships have any paper standing behind them. A lender is not looking for comfort about whether the clause would bind anybody; no lender can get that from a document anyway. A lender is looking for whether somebody thought about the clause, and that much is visible in about ninety seconds.

An analyst covering a listed acquirer reads it to know what the buyer has actually bought. A purchase where 71.93 per cent of the price sits above net assets is a purchase of relationships, and an analyst writing that up wants to say whether those relationships are attached to a system or to four people. The restriction and the non-solicit are the only two places in the public record where that question gets an explicit answer, so they are worth reading even though neither one is a number.

Somebody on the buying side reads it as work still to be done. The clause is the last line of defence and it is a poor one; the real defence is that the four key people want to stay. A restriction is what remains once retention has failed. The same paragraph therefore gets a very different reading from the buying team on day one than from the legal team at signing.

And a household buying a shop reads it exactly the same way, at a scale where the arithmetic fits on a receipt. A tailoring business bought for more than the sewing machines are worth has been bought for the tailor's customers. Whether the paper mentions the tailor by name, and whether it mentions the tailor's two assistants, is the difference between buying a business and buying furniture. The rupees are smaller and not one thing about the question changes.