Portfolio Management: Managing the Whole, Not the List

Portfolio management is the practice of running a set of holdings as one object. Every decision is judged by what it does to the whole rather than by the merits of the holding on its own. Exposure, concentration and liquidity are properties only the whole carries, and a list of separately sensible choices manages none of them.

Underneath that definition sits an idea worth slowing down for. What has been built is not the sum of the reasons for building it. Every holding can be picked well and the thing being run can still be something nobody has understood. Volatility, means and simple ratios are the tools throughout, and each is set out under its own heading.

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. The three sum to Rs 500 crore exactly. Those three shares are policy weightsThe share of a portfolio the holder has decided each asset class should carry. A policy weight is a decision, not an observation, and the actual weights move away from it as prices move., meaning the shape the holder chose in advance, and the actual weights drift away from them between one rebalancing and the next.

What exactly does portfolio management manage?

Ask most people what a portfolio manager does and the answer comes back as choosing what to hold. Choosing what to hold is one of the things the job contains, but it is not the thing the job is. Selection produces candidates. Portfolio management decides what the candidates do to the object they are being added to. The question is a different one and so is the answer.

Think of a kitchen. A meal has properties no ingredient has: balance, timing, whether it can be served hot all at once. A cook choosing ingredients one at a time, each one excellent, does not thereby produce a meal. The same is true here. A holding is judged by what it does to the whole. An excellent holding can therefore be the wrong decision and a dull one the right one, and that single idea sits underneath the whole discipline.

Why would an excellent holding be wrong? Because the whole already carries as much of what that holding brings as the holder decided it could carry. Adding a thirteenth position with the same driver as the twelve already held does not add a thirteenth idea. The thirteenth adds size to one idea. The holding's own merits are real and are also beside the point. The decision was never about the holding.

What separates a portfolio from a collection of holdings?

A collection is a set of separately good decisions that nobody has added up. Every line in it can be defended. Nothing about it is careless. The collection has simply never been treated as one object. The properties that belong to the object have therefore never been measured.

There is a test, and it takes one question. Can anyone in the room state the total exposureHow much of a portfolio depends on one particular thing moving, whether that thing is a market, a currency, an interest rate or a single local condition. to any one thing without opening the list and reading it out? If the answer needs the list, the thing in question is a collection. The test costs nothing and takes one sentence, and almost nobody runs it. So many things called portfolios are collections with better stationery.

The everyday version sits in most households. A household running on one salary makes three sensible savings decisions: a recurring deposit, a small equity plan, and the slow purchase of a home in the town where the employer sits. Each decision is defensible. Read together, all three rest on one employer continuing to pay one salary. The town's demand for housing is that same employer, so the property value rests on it too. Nobody chose that. The concentration assembled itself out of three good choices.

What can only be seen at the level of the whole?

Three things, and they are the reason the discipline exists at all. Total exposure to a shared driver. ConcentrationHow much of a portfolio sits in a small number of positions rather than being spread across many. Concentration is measured against a stated base, and the base has to be named.. And liquidityHow quickly something can be turned into cash without having to move its price much to find a buyer.. None of the three is a property of any holding, so none of the three can be found by reading the holdings, however carefully they are read.

Take the first one. Twelve holdings can each be a genuinely different business and all twelve can still move on one thing. A steel maker, a cement plant, a truck fleet, a tile works, a paint maker, a crane hire operator, a sand supplier, a cable maker, a lift maker, a fittings manufacturer, a housing lender and an estate agent are twelve industries by any classification available. The twelve are also, if they all sell into the same construction cycle in the same state, one position wearing twelve names.

The shared driver is not a fact about any holding and becomes visible only when the twelve are added up, so no holding level review can find it. Each business genuinely has its own customers, its own margins and its own management. The dependence is a property of the set. Asked what it depends on, each holding gives an honest answer that is not the answer required.

How much can a limit on any single holding actually prevent?

The Anantara mandate carries two stated limits of its own, set by its committee rather than by any standard: the equity sleeveThe part of a portfolio held in one asset class. A multi-asset portfolio has an equity sleeve, a fixed income sleeve and so on. must sit between 50 and 70 per cent, and no single holding may exceed 5 per cent of the portfolio. In the stated year the largest holding was 4.6 per cent of the portfolio, or Rs 23 crore, comfortably inside the limit.

Now watch the base do its work. The same holding, measured against the Rs 300 crore equity sleeve rather than against the whole, is 7.7 per cent. Neither number is wrong. The two numbers answer different questions, and an account that slides between them without saying which base it is using has said something untrue about how concentrated the portfolio is. The top ten holdings are Rs 155 crore, or 31.0 per cent of the portfolio and 51.7 per cent of the equity sleeve. Same ten holdings, two figures, twenty points apart.

Here is the part that surprises people. Ten holdings each sitting exactly at the 5 per cent limit would be Rs 25 crore each, or Rs 250 crore together. Rs 250 crore is 50 per cent of the whole portfolio. Would that breach anything? No. The equity band runs to 70 per cent, or Rs 350 crore, so a top ten at half the portfolio sits inside both stated limits with room to spare. The limit caps what any one name can do to the whole. The limit says nothing whatsoever about what the names have in common, and no constraint in portfolio work is more widely misread.

Can a portfolio of easily traded holdings still be slow to move?

Yes, and this is the third property. Whether the whole can be turned into cash at short notice is a fact about the combination and about the timing, not about any single line. Each holding may trade briskly on an ordinary day when only one of them is being sold. Selling all of them at once, to the same set of buyers, on the day those buyers least want to buy, is a different event with a different price.

The street version is a Sunday market. Any one stallholder can sell out by evening. If every stallholder on the street decides to clear stock the same afternoon, the street cannot. The buyers are the same buyers and their appetite has not multiplied. A portfolio of individually liquid holdings can still be slow to move if the same buyers stand behind all of them. Liquidity is measured on the whole and on a stated timeframe for that reason, never holding by holding.

The Anantara mandate carries Rs 50 crore of cash at its policy weight, or 10.0 per cent of the whole. The cash is not idle money in the way it looks on a line by line reading. On a whole level reading it is the part of the portfolio whose timing is certain, and its size is a decision about how fast the rest may need to move.

Who takes the decision, and what is Advisory Portfolio Management?

Two arrangements, and the difference is not a detail. In a discretionaryAn arrangement in which the manager takes and carries out decisions without asking first, inside limits agreed in advance with the holder. arrangement, the kind the Anantara mandate is, the manager decides and executes inside the stated limits. Rukmini Deshpande's committee set the limits; Faiz Ahmad Ansari acts inside them without asking each time.

Advisory Portfolio Management is the other arrangement. The manager recommends and the holder decides each one, so a security moves only when the holder agrees to it. Same analysis, same manager, same limits, and a completely different chain of causation between an idea and a trade.

The consequence is usually skipped and it matters more than the arrangement itself: an advisory record measures the recommendations and the holder's acceptance of them together, so it is not the same measurement as a discretionary record and the two cannot be laid side by side as though the number meant the same thing in both. A discretionary 14.2 per cent reports what a manager's decisions produced. An advisory 14.2 per cent reports what a manager's recommendations produced after passing through a holder who took some and declined others, on their own timing. Two different questions have been answered with the same digits.

What does a stated limit do to the shape of the whole?

A stated limit fixes the shape in advance, before any holding is chosen. Fixing the shape is the whole of its effect, and the effect is larger than it sounds. Once the Anantara mandate says the equity sleeve sits between 50 and 70 per cent, every selection decision that follows is taken inside a box whose walls were built before the first name was considered. The best equity idea in the world cannot take the sleeve to 80 per cent.

Two things about a range are worth separating. A range is not a target. Sitting at the middle of a permitted range is a decision exactly as much as sitting at the edge of it, and calling the middle neutral is how a decision gets taken without anybody noticing they took it. The Anantara policy weight of 60.0 per cent is a chosen point inside 50 to 70, not the range itself and not a default. How a range is arrived at, argued for and written down is covered separately.

Why is beating a benchmark not yet a result?

Here are the stated year's figures for the Anantara portfolio, all belonging to one twelve month period. The portfolio returned 14.2 per cent. The composite benchmarkA stated yardstick a portfolio's result is compared against, built to a published rule so that anybody can reconstruct it., 60 per cent a broad equity index and 40 per cent a broad bond index, returned 12.6 per cent. The risk-free rate for the same period was 6.5 per cent. The excess returnThe difference between what a portfolio returned and what its stated yardstick returned over the same period. is therefore plus 1.6 percentage points, gross, meaning before the cost of running the portfolio has been taken out. Cost comes out below.

Most reporting stops there, and stopping there is the mistake. The portfolio carried a betaA measure of how much a portfolio has tended to move when its yardstick moves. A beta above 1.00 means it has tended to move more, in both directions. of 1.08 against that benchmark, meaning it carried more of the same market than the benchmark did. Some of the 1.6 points is the arithmetic consequence of that and nothing else. Split it and see.

The benchmark returned 6.1 points above the risk-free rate, being 12.6 less 6.5. At a beta of 1.08 the expected figure is 6.5 plus 1.08 times 6.1, or 13.088 per cent. The exposure part of the excess is therefore 0.08 times 6.1, or 0.488 percentage points. The remainder is 14.2 less 13.088, or 1.112 percentage points. The two add back: 0.488 plus 1.112 is 1.600 exactly. Of a headline 1.6 points, about 0.49 points was simply carrying more of the same market and only about 1.11 points was anything else, and those two are produced by completely different things.

There is a shape hiding in that arithmetic. Because the benchmark returned 6.1 points above the risk-free rate, each 0.01 of beta above 1.00 accounts for 0.061 points of the excess. So as beta rises, the exposure part grows in a straight line and the residual shrinks in a straight line, and at a beta of about 1.2623 the exposure part reaches the whole 1.6 points and there is nothing left. Past that, the residual is negative while the headline still reads plus 1.6 points.

Is that 1.6 points gross or net, and does the label change the sign?

Before any cost comes into it, it helps to know what a percentage point of this portfolio is worth. One point of Rs 500 crore is Rs 5,00,00,000/-, so the 1.6 points of gross excess is Rs 8,00,00,000/-, of which the exposure part is Rs 2,44,00,000/- and the residual is Rs 5,56,00,000/-. A point is not a rounding unit at this size. Splitting it is the work, not a refinement to be skipped when the meeting runs long.

Now the word that has been quietly attached to every return figure above. All of them are gross, meaning before the cost of running the portfolio has been taken out. The Anantara mandate carries its own stated commercial terms, a management fee of 1.25 per cent of assets and a performance fee of 15 per cent of the return above a 10 per cent hurdleA stated level of return that has to be cleared before a performance fee applies at all. The fee is not charged below that level., and in the stated year those came to Rs 6.25 crore and Rs 3.15 crore, which is Rs 9.40 crore together. On Rs 500 crore that is 1.88 per cent. Take 1.88 from a gross 14.2 per cent and the holder received 12.32 per cent against a benchmark that returned 12.6 per cent, so a gross excess of plus 1.6 points is a net shortfall of minus 0.28 points over the same twelve months.

Read that twice. Both halves are true at once. The portfolio beat its benchmark and the holder did not. Neither figure is the honest one and neither is the dishonest one. The label is what makes either of them mean anything, so a return with no gross or net beside it is not yet a statement. How the terms of an arrangement are structured, and what any of them may be, is covered separately.

One more comparison falls out of the same two numbers, and it is the one a holder actually asks. The residual left over after the beta split was 1.112 points, measured gross like everything before it, and it is what this arithmetic calls alphaThe part of a result left over once the part explained by the exposure carried has been taken out. Alpha is a remainder, not a measure of anybody's skill.. The cost of running the mandate was 1.88 points. So the cost was larger than the gross excess over the benchmark, and larger than the residual as well. The comparison of cost against residual is the one the holder lives with. The record carries no alternative to set against it, so even that comparison does not settle whether the arrangement was worth keeping.

A last warning, and this one is about arithmetic rather than about money. Elsewhere the same 1.6 gross points is split a second time and differently, into an allocation effect of plus 0.35 points and a selection effect of plus 1.25 points. Set that pair beside the gross 0.488 and 1.112 and the two look as though they nearly agree, and a reader who spots the near miss is tempted to read one as quietly confirming the other.

It cannot. Take the differences inside the same 1.600 gross points: 0.488 less 0.350 is 0.138 points, and 1.250 less 1.112 is also 0.138 points. The repeat of 0.138 is not two quiet agreements but one gap counted twice, forced the moment both pairs are made to total 1.600. The near miss is arithmetic and carries no information about either split. The two answer different questions on different bases, one asking how much of the result was market exposure and the other asking where in the portfolio the result arose, and how the second one is built is covered separately.

How does anybody use this in a room, on a Tuesday?

Three whole level numbers, prepared before the meeting and read before any holding is opened. The three numbers are the practitioner version of everything above, and they are short enough to be done.

An investment committee like Rukmini Deshpande's asks for total exposure to each shared driver, the concentration of the largest holdings stated against a named base, and how quickly the whole could be turned into cash at a stated timeframe. A lender assessing a borrower whose collateral is a portfolio asks the same three. Recovery value on pledged holdings depends on all of them being saleable at once rather than one at a time. An analyst reading a manager's record asks a fourth question before any of the three: what was the beta? A record without the beta cannot be split, and an unsplit record is a headline rather than a result.

A household does the same work with a pen. Each savings decision is written down beside what it depends on, in the plainest words available. If one employer, one town or one interest rate appears on more than half the lines, the total exposure has been found, and found in four minutes without any of the machinery above.

What is portfolio management not?

Portfolio management is not choosing good holdings with extra steps. Selection is a real skill and it is a different one, and a portfolio can be built badly out of excellent selections. The discipline is not a product either, and nothing about it requires a vehicle, a wrapper or a minimum size. Fund vehicles and private structures are delivery arrangements, covered in their own sections, and knowing how one works settles nothing about whether the thing inside it has been managed as a whole.

Exposure, concentration and liquidity are properties of any combination and do not require scale to exist, so the discipline is identical whether the whole is Rs 500 crore or a single household's savings. The Anantara mandate has committees, custodians and a stated benchmark. The household has a pen. Both are managing the same three properties, and both fail in the same way if they only ever look at the list.