The Hedging Policy: What May Be Hedged, and Why

A hedging policy states what may be hedged, with what, against what documented exposure, and how much of it must be covered. Nirjhar Industries Limited, invented, requires at least 60.0 per cent of a committed foreign currency exposure to be hedged and at most 50.0 per cent of a forecast one. A floor on the certain exposure and a ceiling on the uncertain one, and every figure is the group's own.

Almost everybody trips on those two numbers, so they repay a second look. Sixty is a minimum and fifty is a maximum. The floor and the ceiling are not two attempts at one rule, and neither is a typing error somebody should tidy up. Each number answers a different question. Once it is clear which question each one answers, the document stops looking like a list of numbers somebody negotiated and starts looking like an instrument of control. A hedging policy is not a view about where a currency is going; it is a written refusal to hold anything the entity cannot point at a reason for.

Nirjhar Industries Limited, a steel and alloys maker, and Vindhya Commercial Bank Limited, an invented bank, carry every exposure worked below. A forward contract, an interest rate swap and a government security appear as items on a list; what each one is and how it pays belongs to fixed income and derivatives. The subject here is the exposure and the control around it.

What is a hedging policy, and what is it actually for?

A hedging policy is a written document with three parts and one job. The three parts are a list of what the entity may hold, a list of what it may not, and a set of operational caps. The one job is to make sure that every position the treasury carries can be traced back to something the business actually does. A hedging policyA written statement of what may be hedged, with what instruments, against what documented exposure, and how much of the exposure must be covered. at Nirjhar Industries Limited runs to six permitted holdings numbered TP1 to TP6, four prohibitions numbered TQ1 to TQ4, and five treasury limits numbered TL1 to TL5. Fifteen rules, and not one of them is a forecast about the market.

The household version has the same shape. Suppose a family has already booked a hall for a wedding in ninety days at a fixed price, and the caterer wants payment in a currency the household salary does not come in. The family is exposed, and did not choose to be: the exposure arrived attached to a decision already made for a completely different reason. Locking the price now is not a bet on the currency. The lock is a decision to stop the wedding budget from depending on something the family has no view on and no control over. Removing an exposure the family did not choose is the whole idea of hedging, and the policy exists because the same instrument that removes an exposure already held can just as easily create one that never existed.

An instrument that can create an exposure as easily as remove one is the reason a document exists at all. Nothing about a forward contract knows whether the person entering it has a payable behind it. The contract is identical either way. So the control cannot live in the instrument; it has to live in a rule about when the instrument may be used. A hedging policy is therefore written as two lists rather than one, and the second list does most of the work.

What may this group hold, and against what?

Six things, and two of them come with a condition attached that the other four do not. The permitted instrument listThe set of holdings a policy allows, numbered TP1 to TP6 here, which is only half a policy without the prohibitions beside it. reads better as a table than as prose. The condition in the right hand column is the part that matters.

Notice the split. TP1, TP2 and TP3 are places to park cash: the group buys them, holds them and the worst thing that happens is a poor return. TP4 and TP5 are different animals entirely. Each of them can produce a loss far larger than anything the group put in, and each is permitted only against something else. The policy does not permit a forward contract; it permits a forward contract against a documented underlying exposure, and the second half of that sentence is the whole control. TP6 is the odd one out, permitted subject to a rating grade the board sets. The case records that grade as an internal decision and not a market convention.

So the permitted list has already said something before the prohibitions arrive. Two of the six entries carry a condition that points outwards, at a real commercial obligation sitting somewhere else in the group. Remove that condition and TP4 reads simply as permission to enter forward contracts. Permission to enter forward contracts is permission to hold a currency position. One clause is doing all the load bearing, and a reader who skims the list as six bullet points will not see it.

What does the policy refuse outright?

Four things, and they are the mirror image of the list above. The prohibited activity listThe set a policy refuses outright, numbered TQ1 to TQ4 here, and the half that does most of the work. reads: TQ1, no derivative without a documented underlying exposure. TQ2, no position taken to profit from a rate or price view. TQ3, nothing the treasurer cannot value from an independent source. TQ4, no structure whose worst case the board has not seen written down. Every one of the four is about a reason rather than about a size. A rule about reasons is exactly what makes the prohibited list the half that does the work.

TQ1 is the enforcement clause for the condition set out above in TP4 and TP5. Stated positively the requirement is a permission; stated negatively it is a prohibition. The person entering the trade reads the permitted list and the person reviewing it afterwards reads the prohibited one, so policies are written both ways on purpose. The documented underlyingThe requirement that the exposure behind a derivative exists on paper before the derivative is entered into, which is prohibition TQ1 in this group. is therefore tested twice, once before and once after, by two different people. The two sentences those people read say the same thing.

TQ3 and TQ4 are quieter and get skipped in most readings, so hold them for a moment. TQ3 says the treasurer must be able to value the holding from a source that is not the person who sold it. Independent valuation is a control against a position whose loss nobody can see until settlement. TQ4 says the board must have seen the worst case of any structure written down before the group enters it. A written worst case is a control against a position whose loss everybody could have seen and nobody looked for. Between them the two clauses cover the difference between a loss that is hidden and a loss that is merely unexamined. The two are different failures with different fixes.

Why does a certain exposure carry a floor and an uncertain one a ceiling?

Here is where the document earns its keep, and the second rule is worth predicting before it arrives. Treasury limit TL3 at this invented group sets a minimum hedge ratioThe share of an exposure that is covered, stated as a percentage of the exposure and controlled here by a floor on one type and a ceiling on another. of 60.0 per cent on a committed foreign currency exposure. A committed exposureAn obligation the entity is already contracted to, with an amount and a date, so the only open question is how much of it to cover. is one the group is already contracted to: an amount, a counterparty and a date, all settled. The question worth pausing on is what the rule should be on an exposure the group merely expects.

Treasury limit TL4 sets a maximum hedge ratio of 50.0 per cent on a forecast exposureAn expected future exposure that has not been contracted, so it carries a second uncertainty: whether it happens at all.. Lower, and the shape has flipped. Most readers get the direction and miss the shape, and the shape is the finding. A floor states the smallest amount of cover the group will tolerate. A ceiling states the largest amount of cover the group will permit. The floor and the ceiling guard against opposite mistakes, so they point in opposite directions, and no single number can do both jobs.

Worked as two separate questions it becomes obvious. On a committed exposure the amount is known and the date is known, so there is exactly one open question: how much of a known obligation to leave riding on a rate the group has no view on. Every answer to that question is a number between zero and the whole thing. The failure being guarded against is doing nothing, so a floor is the natural shape. On a forecast exposure that question is still there, but a second one sits underneath it: what if the exposure never turns up. A ceiling is not caution about the currency; it is caution about the forecast.

What does the policy actually cost on this group's own payable?

Rupees make it concrete. Nirjhar Trading FZE, a free zone establishment (FZE) that trades for the group overseas, buys in United States dollars (USD) while the group reports in rupees. A locked payable of USD 24 million falls due in 90 days, booked at the group's own contracted rate of Rs 84.00 to the dollar. The contracted rate is an internal figure and not a market rate for anything. Twenty four million at Rs 84.00 is Rs 201.6 crore. The payable is a committed exposure in the exact sense the policy means: the obligation exists, the amount is fixed and the date is fixed, so the only open question is how much of it to cover.

The group models a 5.0 per cent adverse move. Nobody can forecast a currency, so the group sets a scenario of its own and reads the cost off it. Five per cent of Rs 201.6 crore is Rs 10.08 crore, the cost of the move with none of the exposure covered. Apply TL3 and the arithmetic runs in one line.

Rs 4.03 crore reaches the profit and loss account, and it does so because the policy said it could. A loss that breaches nothing is the uncomfortable part. Nobody made an error, nobody breached anything, and the group is fully compliant with its own document while carrying an exposure it could have removed entirely. A floor is not a target, and hedging the whole thing is a different decision that this policy does not require and does not forbid.

Now run the same payable as though it had been a forecast rather than a contract. Suppose the group merely expected to buy USD 24 million of input in 90 days and had signed nothing. TL4 caps the hedge at 50.0 per cent, so at most Rs 100.8 crore may be covered and at least Rs 100.8 crore stays open. On the same 5.0 per cent move the open half costs at least Rs 5.04 crore. The policy has left Rs 1.01 crore more exposed on the less certain of the two, and it has done that on purpose.

What happens to the open cost as cover rises?

The open cost falls in a straight line, and the two policy rules are simply two points on that one line. Once a policy names a number, the temptation is to treat the number as a cliff edge with something different happening on either side. Nothing different happens. The open partThe share of an exposure left uncovered after the policy has been applied, which is a decision the policy makes on purpose rather than an oversight. of the exposure is worth 5.0 per cent of whatever is left, and the arithmetic is Rs 10.08 crore multiplied by one less the hedge ratio. The line is smooth, and the only thing that changes at 50.0 and at 60.0 per cent is whether the group is allowed to be standing there.

Which hedge ratios does this policy leave available to nobody?

Put the two rules on one scale and something falls out that neither rule states. A committed exposure must be hedged at 60.0 per cent or more. A forecast exposure may be hedged at 50.0 per cent or less. So the stretch between 50 and 60 is permitted for neither kind of exposure this policy recognises. A treasurer who writes a ticket at 55 per cent has picked a ratio that is above the ceiling on one type and below the floor on the other, and the document says nothing about it because nobody drafting it looked at the two rules side by side.

Do not overstate the finding. The stretch between 50 and 60 is not a breach, not an error and not a hole anybody has fallen through in this case. The stretch is a property of two rules written by two different people answering two different questions, and the value of noticing it is entirely in the noticing. The cheapest control in any policy is reading it as one document rather than as a set of clauses, and the stretch between 50 and 60 is what that reading produces here. A group that wants the stretch closed can close it in a sentence at the next review; a group that does not want it closed can at least say so deliberately.

What breaks if somebody tidies the two rules into one?

Every policy review throws up at least one proposal to make the document neater, and on a hedging policy the neatening proposal is nearly always the same one. Somebody reads TL3 and TL4 back to back, sees a 60 and a 50 pointing opposite ways, and suggests writing both as a single minimum of 60.0 per cent so the treasury has one number to remember instead of two. The proposal reads like housekeeping. A single minimum is the most expensive sentence anybody could put into that document. The two rules look inconsistent only because the reader has assumed they answer the same question, and they do not.

Follow the proposal through on the forecast side and watch where it lands. Under a single minimum, an expected purchase the group has signed nothing for must now be covered to at least 60.0 per cent, so a ticket goes out against a purchase that exists only in a plan. Then the plan changes, as plans do. The input gets sourced at home instead, or the order shrinks, or the customer the order was for went elsewhere. The purchase never happens. The cover, though, is a live contract with a settlement date, and it does not care in the slightest that the reason for its existence has evaporated.

The group is left holding something with a name, and its own policy already contains that name twice. The live contract is a derivative with no documented underlying exposure, and prohibition TQ1 refuses that outright. The same contract is also a currency position held for no commercial reason, and a position held for no commercial reason is a position on where a rate is going. Prohibition TQ2 refuses that outright. A tidy-up in the limits section has manufactured, by arithmetic alone, exactly the two things the prohibitions section exists to prevent. Nobody set out to take a view on a currency. The document simply stopped telling an obligation apart from an expectation, and that distinction was the only thing holding the two halves of the policy together.

Notice the absences. There is no dishonesty in the chain, no unauthorised trade, no missing signature and nobody exceeding their delegation. Each of the four steps is somebody doing the thing the document told them to do. Most control failures worth learning from have that shape: not a person going round the rules, but a rule quietly instructing a person to do the wrong thing.

What do the other three treasury limits bound?

Two of the five limits carry most of the teaching in this subject and very little of what a treasury actually spends its week on. The remaining three bound things that have nothing to do with how much of an exposure is covered. TL1 bounds exposure to any one bank at Rs 180 crore. TL2 bounds the tenor of any holding at 12 months. TL5 requires Rs 120 crore of liquidity to be held at all times. Every one of the five is the invented group's own board policy, and a hedging policy that only ever talked about hedge sizes would be missing three quarters of what it is for.

Think about what each of the three is guarding. TL1 exists because cover is only as good as the person on the other side of it: put every contract with one bank and a hedge that removes a currency exposure has quietly concentrated a different one. TL2 exists because a treasury holds cash the business will need back, and a holding that matures after the day the money is wanted is not a holding, it is a problem with a maturity date. TL5 exists because a treasury that has arranged everything beautifully and cannot pay a supplier on Tuesday has failed at the only job nobody else in the group can do for it. Read as a set, the five limits bound who the group faces, for how long, and how much of an exposure is left riding on a rate, and that is the whole surface a treasury can get wrong by size rather than by intent.

The right hand column is the column most policies never write, and it is worth reading down on its own. A limit with a number and no stated purpose survives every review by default: nobody can argue against it and nobody can argue for changing it, so it sits there for a decade while the business it was sized for turns into a different business. A limit whose purpose is written beside it can be argued with, and a limit that can be argued with is the only kind that ever gets fixed.

Who approves a hedge, and against what documented exposure?

Two different questions hide inside that one, and the policy answers them in two different places. The first question is who may commit the group at all, a question about delegated authorityThe written permission that lets a named person commit an entity up to a stated size, below which no further approval is needed and above which it is.. The second question is what the commitment has to point at, a question about evidence. A properly authorised person entering a properly sized contract against nothing at all has broken the policy just as thoroughly as an unauthorised one. The group can get the first question right and still fail completely on the second.

On the first question this case records the shape and not the grid. The board writes the policy, and Girish Talwalkar is the group treasurer who works inside it. Prohibition TQ4 gives the board a specific role in one case: no structure whose worst case the board has not seen written down. TQ4 is an approval requirement expressed as a prohibition, and it bites hardest exactly where a treasurer is least able to judge alone. TL1 caps exposure to any one bank at Rs 180 crore, and that cap bounds where the commitments can pile up regardless of who signed them. The case records no delegation table with rupee steps at which one person hands over to another. How authority is delegated in rupee steps is a matter for the group's own delegation table.

On the second question the policy is unusually complete, and it gets there by saying the same thing twice from opposite directions. The person entering the contract reads TP4 and finds a permission: a forward contract is allowed against a documented underlying exposure. The person reviewing it afterwards reads TQ1 and finds a prohibition: no derivative without a documented underlying exposure. One sentence tells the dealer what to attach; the other tells the reviewer what to look for. One test is read forwards by the person taking the risk and backwards by the person checking it, and a policy written only one way is only ever checked by the person who least wants to fail the check.

For the group's own payable the evidence is easy to picture and easy to file: an obligation of USD 24 million falling due in 90 days, with an amount, a counterparty and a date. The paper turns the contract on the other side of it into cover. Take the paper away and the identical contract becomes the only currency position the group has, and no amount of correct authorisation changes what it is. Approval answers who was allowed to do it; documentation answers whether it was a hedge at all, and only one of those two questions has an answer that survives the deal going wrong.

What does this case not record about hedging?

Almost everything, on one side of it. The invented group has a policy, a payable, a ratio and a worked figure. The invented bank has four market risk limits, a balance sheet of Rs 96,000 crore, Rs 84,000 crore of rate sensitive positions and a value change of Rs 840 crore under its own 200 basis point scenario. The 200 basis points are that bank's internal scenario and not anybody's requirement. Against all of that, this case records no hedge at the bank at all: no notional, no ratio, no instrument and no counterparty.

The missing hedge is deliberate and worth sitting with rather than skipping. A case that reconciles across many subjects cannot afford a figure that arrives without a home, and a hedge is never one figure. A hedge is a notional, a direction, a tenor, a counterparty, a valuation source and an effect on two different measures that already have locked answers. Put an invented cover against the bank's banking book position and the Rs 840 crore stops tying to the duration gap that produced it, the earnings figure stops tying to the repricing ladder, and two other subjects in this sequence quietly become wrong. The case therefore does not carry that figure.

There is a teaching point in the asymmetry too, and it is not an accident of the case design. A corporate treasury covers exposures that arrive from operating: a payable in a currency the group does not earn in, a borrowing at a rate the group did not choose. The bank's position is not something that arrived; it is the business. Rate sensitive assets and rate sensitive liabilities are what a bank sells and what it buys, and a bank that covered its interest rate position to nothing would have covered away the reason it exists. So the bank's side of this subject is limits, measurement and committees, and the group's side is a policy and a worked cover, and reading either one as the general case is the error.

Where do the rules come from, and what binds an Indian entity?

Every figure used so far came from a single place. Six permitted holdings, four prohibitions, five limits, two ratios, one contracted rate and one adverse move, and every single one of them was written by the board of one invented company. The mechanism of a hedging policy is jurisdiction free: a permitted list, a prohibited list, a set of caps and a requirement that the exposure exists on paper before the contract does, and that shape would be recognisable to a treasurer anywhere. The numbers filling the shape are not.

In hedging, people quote requirements at each other with great confidence and very little sourcing, so the distinction between the shape and the numbers matters more here than in most subjects. A proportion an Indian company must cover, or a permission a cross-border arrangement does or does not need, gets stated as though it were arithmetic. A threshold printed in a teaching text goes wrong quietly the day the threshold changes, so the binding one has to be read at the authority that sets it.

One last thing sits outside the regulator's door as well as inside it. Whether a cover qualifies for a particular accounting treatment, what has to be documented at the moment it is entered, and what happens in the accounts when it turns out not to be effective, are questions with real answers, and they belong to financial reporting. A treasurer can be entirely inside the hedging policy and still be somewhere unexpected in the accounts, and the two questions are answered by two different documents read by two different people.

So the argument closes where it opened, with two numbers pointing opposite ways. On an obligation the group is already contracted to, the only mistake available is doing too little, so the rule is a floor of 60.0 per cent. On a purchase the group merely expects, a second mistake becomes available, and it is doing too much about something that may never arrive. The rule there is a ceiling of 50.0 per cent. A hedging policy read for the shape of its rules before the size of them reveals what the people who wrote it were actually afraid of.