What an Investment Mandate Is, and What Guidelines Do

An investment mandate is the authority a holder grants to whoever runs the money: what may be bought, within what limits, and with what discretion. Guidelines operationalise it by converting each limit into a figure the manager checks against the portfolio record. The mandate says what is permitted; the guidelines say what that permission is worth today in rupees.

The mandate and the guidelines are two different documents doing two different jobs, and almost every argument about a portfolio limit turns out to be an argument about which of the two somebody was reading. The permission is written in words; the check is done in rupees. The conversion between them is where a limit stops being a sentiment and starts being something a person can be wrong about on a Monday morning.

The running example is the Anantara Multi-Asset Portfolio, an invented discretionary mandate of Rs 500 crore run by Faiz Ahmad Ansari for an invented charitable endowment whose investment committee is chaired by Rukmini Deshpande. Its stated shape is equity 60.0 per cent at Rs 300 crore, fixed income 30.0 per cent at Rs 150 crore and cash 10.0 per cent at Rs 50 crore, summing to Rs 500 crore. Its five written lines are: equity held at a policy weight of 60.0 per cent; equity kept between 50 and 70 per cent; no single holding above 5 per cent of the portfolio; no unlisted holdings; and a minimum credit standing applied to the fixed income sleeve, written as a policy rather than as a rating symbol.

What is an investment mandate, and who grants it?

Start with the ordinary version. A household going away for a fortnight hands a neighbour Rs 20,000/- and a note: buy vegetables and milk, nothing above Rs 500/- in one go, no borrowing, and keep the receipts. The note is a mandate. It is not a shopping list and it is not a description of how the neighbour should shop. A mandate is a grant of authority with a fence around it, and inside the fence the neighbour decides alone.

An investment mandateThe authority one party grants another to act on its money, setting out what may be bought, inside what limits, and how far the other party may decide alone. is that note written for a portfolio. The mandate names who may act, over what pool of money, in what kinds of holding, inside what limits, and with how much discretionHow far the person running the money may decide and act without going back to the holder first. More discretion means fewer conversations before a trade, not fewer limits.. For the Anantara Multi-Asset Portfolio the grant runs from the endowment, through the investment committee that Rukmini Deshpande chairs, to Faiz Ahmad Ansari, over Rs 500 crore, and it is discretionary, meaning he acts and then reports rather than asking and then acting.

Two things a mandate is not, and both mistakes are common enough to be worth naming. A mandate is not a description of a strategy: a strategy is what somebody intends to do inside the permission, and it can change entirely without a word of the mandate moving. A mandate is not a statement of preferences either: a preference is something the holder would like, and a mandate is something the manager may do. A mandate is a permission with edges, and everything outside the edges is refused whatever anybody in the room thinks of it.

An endowment is not the only kind of holder. A private holder handing money to a manager grants a mandate in the same form, answering the same three questions, and nothing in the arithmetic in this guide changes when the holder is one household instead of one endowment. Only the size of the numbers does.

How does a mandate differ from the policy statement it sits inside?

The written policy statement records the purpose of the money, what the holder will accept along the way, and the constraints that follow from both. The mandate answers a narrower question that the statement never touches: who may act on all of that, and how far.

The difference shows when one is broken without the other. Take two changes in turn, and test each of them against both documents. Two documents that fail independently are two documents, however convenient it would be to staple them together.

The two documents carry separate dates for that reason, and the separation is more than filing discipline. One holder can grant several mandates under a single policy statement, and that is the plainest reason the two cannot be the same document. The Anantara record locks one such mandate, the Rs 500 crore multi-asset one, and says nothing about whether others exist.

What is an Investment Guideline, and what does it convert?

An investment guidelineThe working form of a mandate line: the same permission written as a figure that can be worked out from today's portfolio record without anyone having to interpret the wording. is the operational form of a mandate line. Take one sentence of permission, and write beside it the figure that sentence produces when it meets today's portfolio record. Not the sentiment behind the sentence, not a summary of it, but a number or a yes and no that a person with the register in front of them can produce without having to decide anything.

The test is a harsh one, and it is the same test that guideline drafting applies to every written line: could two careful people with the same register arrive at different answers? If they could, the line is not yet a guideline. The guideline is where a mandate becomes enforceable, and a mandate carrying no guidelines is a set of intentions that the manager ends up translating alone. Translating alone is the whole risk. The manager is not being dishonest when he translates; he is doing work nobody else did, in private, and then being measured against it.

Here is the conversion for all five lines. The right-hand column is the manager's actual Monday work.

Three of those five conversions produce a distance and two produce a gate. The difference between a distance and a gate matters more than it looks. A distance can be nearly breached, watched, reported as tightening. A gate cannot: a holding either passes the eligibility test or it does not, and there is no such thing as a holding that is slightly unlisted. Writing both kinds in the same voice hides the fact that they need entirely different monitoring.

The equity line converts into two figures rather than one, and the pair does different work. The policy weight of 60.0 per cent gives Rs 300 crore, the shape the holder chose; the range of 50 to 70 per cent gives Rs 250 crore to Rs 350 crore, the shape the holder will tolerate. The range is a corridor, and a corridor is a very different instruction from a target: reading one as the other is how a drift gets treated as a decision.

Five per cent of what, and struck on which day?

Now the sentence that causes more trouble than any other line in any mandate: no single holding above 5 per cent. The sentence reads as though it has said something exact. Two things are missing, and both are needed before anybody can compute anything.

The first is the denominatorThe total a share is measured against. Change the total and the same holding produces a different percentage without moving at all.. Five per cent of the Rs 500 crore portfolio is Rs 25 crore. Five per cent of the Rs 300 crore equity sleeveThe part of a portfolio held in one asset class, here the equity part. A sleeve has its own total, which is why it can be used as a base and give a different answer. is Rs 15 crore. Rs 25 crore and Rs 15 crore are not two estimates of one figure; they are two different limits, and the largest Anantara holding at Rs 23 crore sits comfortably inside one of them and well outside the other.

The second is the moment the total is struck. A share of a moving total is itself moving, all day, in both the numerator and the denominator at once. A guideline that does not say which day, and at what valuation, the total is taken has left the manager to pick, and picking is exactly what a guideline exists to remove.

The household version is close enough to be uncomfortable. Tell a student that no single expense may exceed a fifth of the month's spending, and the rule sounds firm. Then the month turns out cheap, total spending falls, and a rickshaw fare that was well inside the rule in a normal month is suddenly a breach in a quiet one, without the student having done anything differently. The share moved because the base moved, and a limit written as a share of a moving total will do this every single time.

Work the second reading through and it gets sharper still. If the cap were written against the sleeve, the Rs 23 crore holding would have to fall until it equalled 5 per cent of a sleeve that its own fall was shrinking. The rest of the sleeve is Rs 277 crore, so compliance arrives at Rs 14.58 crore on a sleeve of Rs 291.58 crore. The fall in that holding is 36.6 per cent, not the 34.8 per cent that simply dividing Rs 15 crore by Rs 23 crore suggests. The base matters on the way down exactly as it does on the way up, and by a similar margin.

How much room does 0.4 percentage points actually leave?

The largest holding in the Anantara Multi-Asset Portfolio is 4.6 per cent of the portfolio, or Rs 23 crore, against a cap of Rs 25 crore. A committee report will show two figures for the room left: 0.4 percentage points, and Rs 2 crore. Both are arithmetically true. Both are also close to useless for the only question anybody is actually asking. How much market movement would it take before the holding becomes a problem?

The reason is covered under the written policy statement, where the arithmetic is worked through in full: the holding sits inside its own denominator, so a rise in the holding lifts the cap as well. Rs 2 crore of headroomThe gap between where a portfolio sits and the written line it is measured against, stated as a figure rather than as a general sense of comfort. closes at a rise of 9.15 per cent in the holding, not the 8.70 per cent that dividing Rs 2 crore by Rs 23 crore produces. The cap is reached at Rs 25.11 crore on a total of Rs 502.11 crore.

A guideline writer can run the shortcut in their head. Every rupee the holding gains adds one rupee to the holding and five paise to the cap, so the gap between them closes by only 95 paise. Rs 2 crore of gap therefore needs Rs 2 crore divided by 0.95, or Rs 2.105 crore of gain, and Rs 2.105 crore on Rs 23 crore is 9.15 per cent. A cap written as a share of a total that includes the capped holding always closes at 95 paise in the rupee, and a guideline that reports headroom in points has quietly hidden the other five.

The control below moves only the largest holding. Watch the cap line climb as the column climbs. The climbing cap is the entire reason the crossing arrives so much later than the points figure suggests.

What does the cap permit that it looks like it prevents?

A single holding capA limit on how large any one holding may be, stated as a share of a named total. It binds the largest line and says nothing about the rest. reads like an instruction about spread. It is not. A cap is an instruction about the top of the list, and it is entirely silent about everything underneath. Work out what it permits and the silence becomes obvious.

Rs 25 crore each into Rs 500 crore is twenty holdings, so the Anantara cap taken alone permits the entire Rs 500 crore to sit in twenty names. The same cap equally permits four hundred names at Rs 1.25 crore each, and no reader of that sentence could tell those two portfolios apart.

Bring the equity sleeve into it and the count changes with the sleeve: twelve names at the Rs 300 crore policy weight, fourteen at the Rs 350 crore ceiling, and ten at the Rs 250 crore floor, where ten capped holdings would be the entire equity sleeve.

Set that against what the Anantara Multi-Asset Portfolio actually holds. The top ten equity holdings are Rs 155 crore, or 51.7 per cent of the Rs 300 crore sleeve and 31.0 per cent of the whole portfolio, at an average of Rs 15.50 crore each, or 3.10 per cent of the portfolio each. The cap would permit those same ten names to be Rs 250 crore, or 83.3 per cent of the sleeve and 50.0 per cent of the portfolio.

The permitted top ten is 1.61 times the actual top ten, so the spread this portfolio has was produced by the manager and not by the document. On this measure the cap achieved nothing. The concentrationHow much of a portfolio sits in a small number of holdings rather than being spread widely. It is always measured against a stated total, and the total has to be named. that exists is a choice somebody made well inside the line, and a different manager under the identical mandate could have run something very much heavier.

Look at the rest of the sleeve and the same silence shows up from the other side. The other eighteen names hold Rs 145 crore between them, an average of Rs 8.06 crore each, or 0.94 times what the top ten hold. Across all 28 names the average holding is Rs 10.71 crore, and the cap sits at 2.33 times that average. A limit set at more than twice the typical holding catches one extreme case rather than shaping the portfolio day to day. A limit may reasonably do only that much, provided nobody mistakes it for a spread policy.

Why does one unchanged cap tighten as the sleeve grows?

The effect follows from nothing more than the cap being fixed in rupees while the sleeve is not. The Anantara cap of Rs 25 crore is 10.0 per cent of the equity sleeve when equity sits at its 50 per cent floor of Rs 250 crore. The same cap is 8.3 per cent at the Rs 300 crore policy weight, and 7.1 per cent at the Rs 350 crore ceiling. Not one word of the mandate changed between those three readings.

Read it as a policy and it is genuinely odd. The most any single name may be, as a share of the equity the portfolio is running, is loosest exactly when the portfolio holds the least equity and tightest when it holds the most. Whether the committee intended that is not in the record; what is certain is that one sentence produced three different sleeve policies, depending on where in the corridor the portfolio was sitting on the day somebody looked.

A guideline written against the portfolio and a guideline written against the sleeve are different constraints wearing the same words, and the difference does not sit still. If the committee meant a sleeve constraint, the guideline should say so and the figure should be recomputed as the sleeve moves. If it meant a portfolio constraint, the sleeve reading is a consequence to be aware of rather than a defect. Either is a defensible choice; not choosing is not.

What happens when a limit is crossed by a price rather than a trade?

Somebody buys too much and crosses a cap. Somebody buys nothing at all, the holding rises, and the same cap is crossed. In the report on Monday these look identical: one line, one figure, one number above a limit. In every other respect they are different events, and a guideline that treats them as one has not thought about the harder of the two.

A passive breachA limit crossed without anybody trading, because prices moved the holding or the total. Nobody decided anything, and yet the written line has been crossed. is the second kind. Nobody in the building acted. The market did the crossing. The wording of the mandate on a passive breach determines what the manager must now do, and there are only three settings available.

Forbidding both kinds commits the manager to selling into a rising holding on a fixed schedule, every time the market pushes a good position through the line. Forbidding both is a real policy with a turnover cost no return figure ever shows, and a committee should choose it rather than discover it. Forbidding the purchase only makes the passive crossing something to report and cure inside a stated period. Forbidding neither leaves no cap at all: there is a number in a document that nothing ever obliges anyone to act on. A cap that names neither kind of crossing is not a limit, it is a sentiment, and the difference only becomes visible on the day it is crossed.

Which setting does the Anantara mandate use? The record does not say. Anybody who cannot separate what a record locks from what it leaves open will eventually invent the second and quote it as the first, so an absence is worth naming rather than filling in.

Who may change a mandate, and what does a breach oblige?

Two events that get confused, expensively. A variation is the holder changing what is permitted. A breach is the portfolio being outside what is permitted. The first is a decision and travels one route; the second is a fact and travels another.

A variation runs from the holder, through the investment committee Rukmini Deshpande chairs, in writing and dated, and then to the manager, who acts on the new permission from that date forward. A manager who can move the fence is not being constrained by it, so Faiz Ahmad Ansari cannot vary his own mandate in any direction, including tightening it. He can report, and must.

Rewriting a line after the portfolio has crossed it does not turn the breach into a variation; it produces a variation dated after a breach, and the record now carries both. The distinction between a variation and a breach is the entire reason the dates exist. A committee that quietly re-cuts a limit to fit the position has not fixed anything: it has recorded, permanently, that the position came first and the permission second.

How does anybody use this in a room, on a Monday morning?

Concretely, and it takes one sheet. The manager arrives with the guideline figures, the current reading against each, and one line most sheets leave out. The missing line is the honest headroom: not the gap in points, but the price movement it would take to close it. Five rows, no commentary. Anything that needs a paragraph belongs elsewhere.

The committee reads the same sheet with a different question in mind. The status column already answers whether the manager complied. Rukmini Deshpande is checking something else: whether the lines still say what the endowment meant. A guideline that has never once come close to binding may be doing no work at all, and a guideline that binds every quarter may be describing a portfolio the holder no longer wants. Both are variation questions, and both belong to the holder rather than the manager.

An analyst reading somebody else's mandate from the outside uses it differently again, as a bound on what the record can possibly mean. If a manager reports a strong year and the mandate permits twenty names, the analyst knows the result is consistent with a very concentrated portfolio and cannot tell from the mandate alone whether it was one. A lender looking at a borrower's investment policy does the same in reverse, reading the guidelines as the worst case the borrower is permitted to run rather than the case it runs today.

The household version is smaller but not different. The rule a household has given itself goes in one column, and beside it the rupee figure it produces this month and the total that figure was measured against. Two columns, once. The base was never named, so most people find on the first attempt that the rule they thought they had cannot be checked.