Writing an Investment Policy Statement, Step by Step
Writing an Investment Policy Statement, Step by Step
Writing an investment policy statement runs in a fixed order: the purpose of the money and the date it is needed, then the return objective and the risk objective together, then the constraints, then who may change any of it. Every line has to reduce to something a person can check today without asking the drafter what it meant. A line that cannot be checked cannot govern.
The drafter is writing a document that will still be read after everyone in the room has moved on. The whole design problem starts there. The people who will use it are a manager reading it on a busy morning, a committee reading it after a bad quarter, and an outsider reading it cold two years from now. None of them can ask the drafter what was meant.
So the drafting runs as a sequence, and the sequence is not a matter of taste. Each section is derived from the one above it. A document that opens with an allocation has begun in the middle. Everything below is the order, one step at a time, worked on the Anantara Multi-Asset Portfolio, an invented Rs 500 crore mandate run for an invented charitable endowment.
Where does the drafting actually start?
Drafting starts with two sentences that nothing else in the document can supply: the purpose of the money, and the date it is needed. Not the size of the portfolio, not the appetite for risk in the abstract, and certainly not a table of weights.
The household case comes before the endowment. Two neighbours have the same savings to the rupee. One is putting a daughter through a five year course starting next June. The other has no call on the money at all and would like to leave it to a trust. Nobody would expect those two to write down the same instructions, and nobody would think the difference had anything to do with how much they had. The purpose and the date are the only inputs to a policy statement that cannot be derived from anything else. Two holders with identical wealth end up with genuinely different documents.
For the Anantara Multi-Asset Portfolio the two sentences are that the money supports the endowment's spending in perpetuity, and that a fixed share of it is drawn each year with no single large call foreseen. The two sentences go down before anything else. The rest of the document is an argument that follows from it, and a drafter who skips this step ends up defending weights that cannot be traced back to a reason.
Two holders have identical wealth and end up with different investment policy statements. What made them different?
Why do the two objectives have to sit together?
Step two writes the return objectiveThe statement of what the money is being asked to earn, expressed against something, such as a rate of spending or a stated comparison. and the risk objectiveThe statement of how much movement in value the holder is prepared to live with while the return is being pursued.. Both of them. In the same sentence, or in two sentences that touch.
Putting them together looks like a formatting preference and is not. A return objective separated from its risk objective by three sheets of paper gets read on its own, and a return objective read on its own is a target that can be hit by simply taking more exposure. Physical placement changes how a document is read. Put them side by side and neither one can be quoted without the other. Keeping the two objectives adjacent is the entire reason for the rule.
For the Anantara Multi-Asset Portfolio the pair reads as one line: the portfolio pursues a long term return sufficient to sustain the endowment's annual draw and its purchasing power, and it does so accepting the movement in value that a majority holding in equities produces over a full cycle. One sentence, two objectives, and nothing separating them. How each objective should be worded for a particular holder is settled under the return objective and the risk objective themselves.
Where do the constraint lines come from?
Step three writes the constraints. A constraintA written limit on what may be held or done inside the portfolio, as distinct from an objective, which states what the portfolio is trying to achieve. is not invented at the desk. A constraint is collected, by asking four questions in turn and writing down whatever comes back.
The four questions are what the law requires of this holder, what cash is needed and on what date, what the holder's tax position does to a realised gain, and what is true of this holder and nobody else. Ask them in that order, write the answers, and stop. A borrowed list produces lines nobody needed and quietly leaves out the one that mattered, so the drafter writes constraints from the four sources instead.
The pull to explain any of them at this step is worth resisting. A drafting step that starts describing what a liquidity requirement is has stopped drafting and started teaching, and those explanations are covered separately. At this step the four questions are a collection instrument, nothing more.
What test does every drafted line have to pass?
Step four is the centre of the drafting sequence. At this point the draft is a list of sentences and nothing more. Before any of them is allowed to govern anything, each one gets handed to a person who was not in the room, along with the portfolio record, and that person is asked one question. Can this line be shown as satisfied today, using the portfolio record alone?
The question has a name, the checkability testA single question put to each drafted line: can an outsider, holding only the portfolio record, state today whether the line is satisfied.. A line that survives it governs. A line that needs somebody to explain what the drafter had in mind does not govern, however sensible the idea behind it is, and however carefully the sentence was built.
Notice what is being tested. Not the idea. The writing. A committee can be completely right that concentration is the thing to watch and still produce a sentence that settles no argument. The test does not ask whether the worry was real. The test asks whether the words can be resolved against a record, by somebody with no memory of the meeting.
A drafted line reads: the portfolio shall avoid excessive concentration. Can it govern?
What does a surviving line get turned into?
Step five converts. Every line that passed the test is rewritten as the thing the manager will actually look at on a screen. A percentage of the portfolio becomes a rupee figure. A range becomes two rupee figures. An exclusion becomes a yes or a no, asked once per holding.
A conversion left to whoever reads it later is a conversion two people will perform differently, so the conversion is done at drafting time, inside the document. The arithmetic is trivial, which is exactly why it gets skipped. On a Rs 500 crore portfolio, a policy weightThe share of the portfolio a document sets for an asset class as its standing position, before any drift caused by prices between one rebalancing and the next. of 60.0 per cent in equities is Rs 300 crore, being Rs 3,00,00,00,000/-. The conversion takes four seconds and removes an entire category of later argument.
There is a second reason, and it is about the reader rather than the arithmetic. A percentage has to be applied to a base before it means anything, and a document that leaves the base unstated invites two readers to pick different bases in good faith. Writing the rupee figure fixes the base in ink.
Convert a 5 per cent single holding cap on the Rs 500 crore Anantara Multi-Asset Portfolio into the figure a manager checks.
What happens when the five lines are actually tested?
The Anantara Multi-Asset Portfolio is governed by five lines, drafted by a committee chaired by Rukmini Deshpande and run by Faiz Ahmad Ansari. Here they are, run through the checkability test one at a time, with the conversion done.
| The drafted line | What a manager checks | Passes |
|---|---|---|
| Policy weights of equity 60.0 per cent, fixed income 30.0 per cent and cash 10.0 per cent | Rs 300 crore, Rs 150 crore and Rs 50 crore, summing to Rs 500 crore | Yes |
| Equity between 50 and 70 per cent | Rs 250 crore and Rs 350 crore, two edges | Yes |
| No single holding above 5 per cent of the portfolio | Rs 25 crore, tested holding by holding | Yes |
| No unlisted holdings | A yes or a no on each holding | Yes |
| A minimum credit standing on the fixed income sleeve | Nothing at all, until the document defines the standing | No |
| Five lines | Four convert without further work | 4 of 5 |
Four of the five lines pass unaided and the fifth converts to nothing, and presenting this mandate as uniformly well drafted would teach a reader to stop looking. So look at the fifth. Why would a committee that got four lines right write the fifth one loosely?
The honest answer is that the alternative was worse. Write the line as a rating symbol and the document has borrowed somebody else's scale, along with whatever methodology sits behind it. If that scale is revised, or the meaning of one of its symbols shifts, then what the mandate permits has changed and nobody in the room decided anything. The document quietly means something new on a morning when no committee met.
A line written as a borrowed symbol lets a governing document change meaning without anybody having decided, a worse failure than a line that needs one more paragraph. So the resolution is neither the symbol nor the vague phrase: it is that the document supplies its own definition of the standing, inside itself, where the committee can be held to it. Rukmini Deshpande's committee accepted that as a drafting cost rather than an error, and the count stays four of five until that definition is written.
Four of the five drafted lines convert straight to a number. Which one does not, and why was it written that way anyway?
Is a tighter range always better discipline?
How tight should a line be written?
Step six is where drafters reach for a principle and there is arithmetic available instead. The question is how wide to write a tolerance rangeThe corridor a document allows a weight to sit inside, stated as two edges around the standing weight, rather than a single point., and the honest way to answer it is to compute what each candidate range hands over.
On the Anantara Multi-Asset Portfolio, one percentage point of the portfolio is Rs 5 crore, being Rs 5,00,00,000/-. The mandate's equity corridor of 50 to 70 per cent is twenty points wide, so it gives Faiz Ahmad Ansari Rs 100 crore of equity exposure to move without asking anyone. Narrow it to 55 to 65 per cent and the same portfolio, unchanged in every other respect, gives Rs 50 crore, exactly half. Narrow it again to 58 to 62 per cent and it gives Rs 20 crore, one fifth of the original.
The width of a range is a rupee figure, and a committee that writes a range without computing that figure has decided how much authority to hand over without knowing how much it was. That is the finding, and it cuts both ways. A tighter range is not automatically better discipline: it takes back delegated discretionThe room a document gives the manager to act inside without returning to the committee for permission each time., and prices move the weights on their own, so it also makes a breach more likely on days when nobody traded at all. The arithmetic is fixed by the size of the portfolio. The width itself is the committee's decision, and a different committee will write a different one.
A committee narrows the equity range from 50 to 70 per cent down to 55 to 65 per cent. What have they changed?
The discretion meter
One control, one consequence. Widen and narrow the equity corridor around the fixed policy weight of 60.0 per cent and watch what the drafting choice hands over. The portfolio stays at Rs 500 crore throughout and the policy weight never moves.
A range 20 points wide runs from 50.0 to 70.0 per cent, which is Rs 250 crore to Rs 350 crore, and it lets the manager move Rs 100 crore of equity exposure without asking the committee.
Educational illustration. Widen it and read the rupees. No width is a right answer, and the marks at 4, 10 and 20 points are simply the three cases computed above.
Who signs it, and what may change it?
Step seven closes the document. Record who signed it and on what date, and then write the change routeThe written path by which a line in the document may be altered: who may propose it, who must agree, and how the change is recorded.: who may propose an amendment, who must agree to it, and where the amendment is written down.
Step seven reads like administration and is not. A document with no change route gets changed informally, and an informally changed limit can never afterwards be told apart from a limit that was simply crossed. Picture the conversation two years out. The equity weight sat at 72 per cent for a quarter. Was the corridor widened by agreement, or was it breached and nobody wrote it down? With a change route the record answers. Without one, the two possibilities are indistinguishable, and the document has lost the one thing it existed to produce.
For the Anantara Multi-Asset Portfolio the closing line records Rukmini Deshpande's signature as chair of the investment committee, the date of adoption, and that any change to a governing line requires a written amendment carried at a meeting of the committee and attached to the document itself.
The document says nothing about who may change it. What happens?
The error that gets made, and what it costs
A drafter writes a line that reads beautifully and cannot be checked: the portfolio shall maintain adequate liquidity and avoid excessive concentration. Both halves sound like limits. Neither one is. Adequate and excessive are not quantities that appear anywhere in a portfolio record, so nobody can say today whether the Anantara Multi-Asset Portfolio satisfies either.
The cost arrives late and arrives badly. When a position is finally questioned, the manager and the committee each read the sentence in their own favour, and there is no fact available that settles it between them. Worse, the sentence has consumed the space where the real line would have gone, so the committee believes it holds a limit on concentration when what it holds is a mood.
The fix is the test itself: hand every line to somebody who was not in the room, ask them to resolve it against the record, and convert or cut whatever fails. A good idea written untestably governs nothing at all.
How the finished document actually gets used
None of this is drafting for its own sake, and that shows in what happens to the document afterwards. A monitoring report for the Anantara Multi-Asset Portfolio is built by putting each converted line in a column and the current figure beside it: Rs 25 crore against the largest holding, Rs 250 crore and Rs 350 crore against the equity sleeve. Somebody who has never met the committee can produce that report, and producing it is what the drafting was for.
An analyst reviewing the mandate from outside does the same thing in reverse. The analyst reads the five lines, converts them, and checks whether the portfolio sits inside. Where a line does not convert, the analyst cannot review it and says so, and a drafting weakness surfaces long before it causes an argument.
The test of a policy statement is whether a stranger holding only the document and the portfolio record can produce the compliance report, and every step above exists to make that possible. A household writing its own version has the same test available: a rule that cannot be checked at the end of the month without reinterpreting it is not yet a rule.
What single question decides whether a drafted line governs?
Where the requirements outside the room are published
A written mandate may have to satisfy requirements that were not set in the drafting room. Such requirements are published by the Securities and Exchange Board of India at sebi.gov.in and, for retirement arrangements, by the Pension Fund Regulatory and Development Authority at pfrda.org.in. The current position should be confirmed at those sources. The drafting sequence above is about how a document is written, and it holds wherever the mandate is being written.
References
| Source | What it is named for | Where |
|---|---|---|
| Securities and Exchange Board of India | Named as the publisher of requirements a written mandate may have to satisfy. | sebi.gov.in |
| Pension Fund Regulatory and Development Authority | Named as the publisher of requirements applying to retirement arrangements. | pfrda.org.in |
The Anantara Multi-Asset Portfolio, its endowment holder, Rukmini Deshpande and Faiz Ahmad Ansari are invented.
Educational material. Not advice on any investment, tax, budget or market position.
