How to Create a Portfolio Factsheet a Holder Can Use

A portfolio factsheet is the single sheet the holder actually receives, and it works only when every figure on it carries its window, its base and whether it was struck before or after the cost of delivery. The Anantara Multi-Asset Portfolio's stated twelve month period reads 14.2 per cent gross, 12.32 per cent net of an invented 1.88 per cent of fees, against a benchmark of 12.6 per cent.

Nothing on the sheet being built is a new calculation. The return, the volatility, the beta, the ratios and the fee arithmetic are all settled elsewhere, and assembling them is its own skill. A factsheetThe short periodic report a holder receives about a portfolio, carrying its figures for one stated period in a fixed set of fields. is a fixed set of fields in a fixed order, and the order is load bearing rather than tidy. Each field exists because a figure below it cannot be read without it. Move the window to the bottom and every return above it becomes unreadable.

Ten fields run in order, and each one takes something from elsewhere, prints something on the sheet, and leaves something open. How each measure is arrived at is covered separately. Every figure below belongs to the Anantara Multi-Asset Portfolio, an invented Rs 500 crore discretionary mandate run for an invented charitable endowment, over one stated twelve month period.

Who is a factsheet actually written for?

Saying the sheet is for the client is true and it does not settle what to print. Three different people pick the sheet up, and each of them is trying to do something the others are not.

The holder's investment committee, chaired on the Anantara portfolio by Rukmini Deshpande, reads it once the period has closed and asks one question: what did the money actually do. The endowment spends out of the portfolio. The question is about what arrived rather than what was earned before costs.

The person running the mandate, Faiz Ahmad Ansari, reads it as the record of a period that has finished. He cannot change a figure on it and cannot argue with a field that is missing, so the sheet has to be complete rather than favourable. A sheet designed to be defended is a different document from a sheet designed to be checked, and only the second one is a factsheet.

The third reader is the one most sheets forget: somebody arriving cold two years later with this sheet and nothing else, no minutes and nobody to ask. Every field has to survive that reader, and that is the test which decides whether a figure gets its base printed beside it.

Three readers decide the drafting. A field earns its place not by being interesting, but by carrying something the third reader could not resolve alone, and the sheet prints every field that passes that test.

Which field comes first, and why is it not a formality?

Field one identifies the portfolio and states the window: the portfolio's name, its holder, the period covered with a start and an end, and the date the sheet was produced. Four short lines, and every one earns its place.

The window gets treated as housekeeping and is not. A return figure with no window attached can be quoted over any window at all afterwards, and nothing on the sheet contradicts the requote. Say 14.2 per cent with no period beside it and the next person can attach it to a quarter or to three years.

A street vendor reports takings of Rs 40,000/-. Nothing is known yet. Rs 40,000/- in a day is one business, Rs 40,000/- in a month is a different one, and Rs 40,000/- across the wedding season is a third. The number never changes and the meaning changes three times. Nobody is lying and nobody can read the figure.

On the Anantara sheet the window is one stated twelve month period with its start and its end printed, and every return, volatility, ratio and count below belongs to it. The record holds one period only. The sheet does not annualise that period, stretch it, or sit it beside another.

What has to be said before any return figure appears?

Field two is one line long and it changes the meaning of every figure below it. Field two states the return basisThe statement of which costs have already been taken out of a reported return before it is printed.: whether the returns are struck before or after fees, and before or after dealing costs. The basis is not a caveat in small type at the foot, and gross of feesA return figure struck before the cost of running the arrangement has been deducted from it. and net of feesA return figure struck after the cost of running the arrangement has been deducted from it. are two different figures about the same year.

The Anantara record states that returns are struck before fees unless a line says otherwise, and says nothing about dealing costs. The silence about dealing costs is a genuine gap, and the sheet prints the gap as a sentence rather than choosing a convenient reading of it. The gap could be resolved either way at the desk and nobody would catch it. Being uncatchable is exactly the reason not to. Assuming the friendlier reading makes a decision on the holder's behalf without saying a decision was available. Portfolio turnover leaves this basis open for the same reason.

How does a benchmark line get written so it can be checked?

Field three describes the benchmark rather than naming it. The Anantara benchmark is a composite of 60 per cent a broad equity measure and 40 per cent a broad bond measure, and it returned 12.6 per cent over the same stated twelve month period.

Three things go on the line, each of them there for the third reader: the construction, the period it was struck over, and the figure itself. A benchmark line without its construction cannot be checked by the person reading it. The comparison then becomes something the reader is asked to take on trust.

How the measure is built and maintained belongs to the organisation that publishes it, and the exchanges at nseindia.com and bseindia.com are where construction rules for Indian market measures are published. The sheet points and does not paraphrase.

Two smaller disciplines sit inside this field. A 60 and 40 composite and a 50 and 50 composite give different comparisons, and the returns alone cannot tell them apart. So the composite weights are printed rather than implied. And the benchmark is never shortened to the index later on.

Why does the sheet carry three return lines rather than one?

Field four is the field this sequence turns on, and the rule is about how many lines get printed. Three, in order: the gross return, the fee deduction, and the net return. Then the excess against the benchmark, labelled on both bases.

On the Anantara sheet those lines read 14.2 per cent gross, a fee deduction of 1.88 per cent of assets, and 12.32 per cent net, against a benchmark of 12.6 per cent over the same window. So the sheet prints a gross excess of plus 1.6 percentage points and a net excess of minus 0.28 percentage points.

Plus 1.6 and minus 0.28 describe the same portfolio, the same year and the same benchmark, and differ only in whether the cost of delivery has been taken out. So no excess return goes on the sheet without the word gross or the word net in the same line. Both are correct, and neither of them is the excess return.

Percentage points slide off the eye, so the rupee version is worth printing too. A gross excess of 1.6 points on Rs 500 crore is Rs 8,00,00,000/-, the fees were Rs 9,40,00,000/-, and the gap of Rs 1,40,00,000/- is the same minus 0.28 points in money the holder can picture.

What has to be printed beside each risk figure?

Field five is where a factsheet most often prints figures that look precise and mean nothing. Every risk figure needs a companion beside it, and the companion is what makes the figure comparable with anything at all.

Volatility goes on the line with the benchmark's, 11.8 per cent against 10.4 per cent over the same window. Beta, 1.08, goes with what it was measured against, that same composite. Tracking errorA single number describing how far a portfolio's period returns tend to sit away from the benchmark's over the same period. of 3.7 per cent is printed against that named comparison rather than on its own.

Any ratio gets the risk-free rate printed beside it, and the Anantara sheet uses 6.5 per cent. A ratio without its risk-free rate is not comparable with anything and should not be printed at all. The reader cannot tell what rate it was struck against. The ratio of excess return to total volatility is William F. Sharpe's, and the name goes on the sheet where the measure appears.

Then the drawdown. The Anantara portfolio's worst fall peak to troughMeasured from the highest point a value reached inside a stated window down to the lowest point it reached after that high, within the same window. inside the stated year was 9.7 per cent, against 8.1 per cent for the composite, and the figure travels with its window every time. A different window gives a different number.

The ratio question is not a trick and the answer is on the sheet. Struck on the gross return the Anantara ratio is 14.2 less 6.5 over 11.8, or 0.653, against the composite's 12.6 less 6.5 over 10.4, or 0.587. Struck on the net return the same ratio runs 12.32 less 6.5 over 11.8, or 0.493. Nothing changed between the two readings but the basis. A ranking that reverses on a change of basis alone is the strongest argument there is for printing the basis in field two rather than at the foot of the sheet. The sheet prints all three and picks no favourite.

The same doubling applies to the information ratioThe excess return over a benchmark divided by the tracking error against that same benchmark, both over the same stated period.: 1.6 over 3.7 is 0.43 on the gross excess, and minus 0.28 over 3.7 is minus 0.08 on the net. The sheet prints the pair and never one of them alone.

How does the constraint section get recorded?

Field six is the constraint registerThe section of a report that lists each written limit on the portfolio, the figure it was tested against, and whether it held for the period., and it has four columns rather than two. The constraint, the figure it was tested against, the base that figure was struck on, and the result with its date if anything failed.

Most sheets carry two of those four: the constraint and a tick. A constraint and a tick are enough to be reassuring and not enough to be checked. For a document meant to be checked by somebody who was not there, that is the wrong way round.

The base column is the one that goes missing, and on this portfolio it is the one that matters. The largest holding on the Anantara sheet is Rs 23,00,00,000/-, or 4.6 per cent of the Rs 500 crore portfolio and 7.7 per cent of the Rs 300 crore equity sleeve. The cap is written against the portfolio at 5 per cent, so the holding is inside it.

Neither percentage is wrong and they answer different questions, so a pass recorded without its base cannot be checked and will be re-derived from scratch under pressure at the worst possible moment. Print the base in the same row, every row, and the argument never starts.

The same discipline runs through the concentration lines beside the constraints. The equity sleeve holds 28 names, and the top ten are Rs 1,55,00,00,000/-, or 31.0 per cent of the portfolio and 51.7 per cent of the sleeve. The two readings answer different questions, so both go on with their bases attached.

How are the fees shown, and against what?

Field seven prints the cost of delivery: the total in rupees, the total as a share of assets, and the arrangement restated so a reader can rebuild both without asking anybody.

The Anantara mandate carries a management fee of 1.25 per cent of assets, or Rs 6,25,00,000/- on Rs 500 crore, and a performance fee of 15 per cent of the return above a 10 per cent hurdle. The stated year returned 14.2 per cent, so the excess over that hurdle was 4.2 points, or Rs 21,00,00,000/-, and 15 per cent of that is Rs 3,15,00,000/-. Total Rs 9,40,00,000/-, or 1.88 per cent of assets.

Readers convert in different directions, and the one who cannot convert simply does not. So printing 1.88 per cent without printing Rs 9,40,00,000/-, or the rupee figure without the percentage, halves the disclosure. A household knows it already: a mobile plan quoted at eleven per cent more this year means less than the same plan quoted as Rs 90/- more a month.

The terms get restated for a separate reason. A reader who can see the 1.25 per cent, the 15 per cent, the 10 per cent hurdle and the Rs 500 crore base can rebuild Rs 9,40,00,000/- from scratch; a reader handed only the total has to believe it. The fee terms belong to this mandate alone. No market sets them and no regulator sets them.

There is a second comparison the sheet can print and a third it must refuse. The sheet can print the fees against the gross excess of Rs 8,00,00,000/- and against the alpha for the stated year, the residual of 1.11 percentage points at a beta of 1.08, or Rs 5,55,00,000/-. The residual measure is Michael C. Jensen's, credited to him where it appears.

The sheet cannot print a verdict. Whether the arrangement was worth having depends on what an alternative would have returned and what it would have cost, and this record contains no alternative at all, so the sheet states the comparison and stops.

What turns a count into a figure?

Field eight carries activity and stewardship, and both arrive as counts that are useless until something is printed beside them. Counts are where a sheet most often looks informative and conveys nothing.

Turnover first. The Anantara portfolio turned over 34 per cent for the stated period, about Rs 1,70,00,00,000/- of trading on Rs 500 crore. The percentage alone leaves a reader to picture what a third of a portfolio being replaced amounts to, and the rupees alone hide how big it is relative to the whole. Both go on.

Then the votes. The mandate voted on 214 resolutions during the stated period and voted against the board's recommendation on 19 of them. Nineteen. Whether that is a lot cannot be said, by the reader or by the holder, until the denominatorThe figure a count is divided by to turn it into a share, printed beside the count so the share can be checked. is printed beside it. Nineteen out of 214 is 8.9 per cent.

A count without its denominator is not a figure, so the sheet computes the share and prints the two numbers it came from. Then the sheet stops. The mix of resolutions is not in the record, so even 8.9 per cent supports nothing about how the mandate conducted itself.

How is something the record does not contain written down?

Field nine is the field almost no factsheet carries, and it is what most separates an honest sheet from a polished one. Anything the sheet cannot say gets written down as a sentence saying so, with the reason attached.

The instinct is to leave the field blank. Resist it: a blank does not read as unknown. A blank reads as nil, or as not applicable, or as nothing worth mentioning, so a blank is a hole while a named absenceA sentence printed in place of a figure that does not exist, saying what is missing and why, instead of an empty space. is a disclosure.

A rent receipt with the maintenance line left empty splits its readers: half conclude no maintenance was charged, half conclude somebody forgot, and both halves are guessing. Write down that maintenance was not billed this month and there is nothing left to guess about.

The Anantara sheet carries five of these, and every one is a genuine gap in the record rather than a flourish.

What never appears on the sheet at all?

Field ten is a refusal rather than a field, written before the sheet goes out rather than argued about afterwards. Five things stay off the sheet permanently. No forecast of what the portfolio will return. No target for any figure on the sheet. No claim that the stated period establishes anything about a later period. No comparison with a portfolio the reader could not have held. And no conclusion about whether the year was good. A conclusion is a judgement rather than a report.

A factsheet reports one period and stops. Everything beyond that is somebody's opinion wearing a report's formatting, and a report's formatting is the most persuasive form an opinion can take. The refusal is structural for that reason: an opinion set in the same typeface as a measured figure reads as measured.

The sheet also declines to split its own excess. There is more than one way to divide 1.6 points and the splits answer different questions on different bases, so printing one without naming its question would assert something the record has not settled. Splitting the excess belongs to performance monitoring.

What must never be a step in building the sheet

Three moves must never enter this procedure, however reasonable they look at the desk. Never recompute a figure that is already settled elsewhere. A second derivation is a second answer nobody downstream can reconcile. Never resolve a gap in the record so the sheet reads cleanly. Field nine exists to prevent exactly that. And never choose between two correct readings of the same figure. Choosing is what field four and field six were built to stop.

What does the finished sheet look like end to end?

Here is the whole artefact, assembled from the ten fields in order. Every line on it belongs to the one stated twelve month period.

Taken together, the three parts show how much of the sheet is spent on companions rather than headline figures. Companions are the difference between a sheet a stranger can check and a sheet a stranger has to believe.