Tax Constraints in a Mandate: Whose Position Counts

A tax constraint reaches a portfolio through the holder rather than through the holdings: the same positions in the same weights leave a different amount behind for two holders in different positions. The constraint bites at the moment a gain is realised. How much the portfolio trades sets the size of the reduction. The treatment itself comes from the tax authority.

Naming an outside authority is not a hedge. The dependence on a body outside the portfolio is the shape of the whole subject. Every other constraint in a governing document can be written down, checked and audited from inside the room: an equity band of 50 to 70 per cent either holds or it does not, and anyone with the holdings list can say which. A tax constraintA limit on a portfolio that arises from what a realised gain costs the person or institution that holds it, rather than from anything written in the portfolio itself. cannot be checked that way. Half of what determines it sits outside the portfolio entirely, and the other half changes when somebody presses the sell button.

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 policy weights are 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. Over one stated twelve month period it returned 14.2 per cent gross against a composite benchmark of 60 per cent a broad equity index and 40 per cent a broad bond index that returned 12.6 per cent, and it replaced about a third of itself along the way.

Before any of that becomes useful, the gaps in the invented record matter more than its contents. The invented Anantara record locks the return, the benchmark, the beta, the traded value and the turnoverThe share of a portfolio replaced over a stated period, measured from traded value rather than from the number of decisions taken.. The record locks nothing about what the endowment pays on a gain, and nothing about how much of the return was crystallised rather than merely showing on a valuation. A relationship between the terms can therefore be worked out, and a result cannot.

All three absences sit on the same side of a line. The split is not an accident of an invented record. Portfolio facts and holder facts are settled by different people, in different places, for different reasons.

What is a tax constraint, and where does it come from?

Start with the everyday version. The finance version is the same shape wearing a suit. Two neighbours buy the same scooter from the same dealer on the same morning at the same price. One rides it to a salaried job; the other runs a delivery round on it and books the running cost against the earnings from that round. Same machine, same price, same street. The scooter finally costs each of them a different amount, and nothing learned by examining the scooter would tell which is which.

A tax constraint on a portfolio works exactly like that. The constraint is not a property of the shares, the bonds or the cash. A tax constraint is a property of the arrangement between the person or institution holding them and the authority that publishes what a realised gain does. Two mandates holding identical positions in identical weights can carry completely different tax constraints, and neither manager can see the difference anywhere in the portfolio record.

The sources of constraint on a governing document are for that reason usually listed with the tax position as a separate line rather than folded into the others. Four questions with four different origins sit behind them: what the law asks of the holder, what cash the holder needs and by when, what a realised gain costs the holder, and what is true of this holder and nobody else. The third one is the only one whose answer is invisible from the portfolio side.

There is a second consequence, and it is the reason the subject needs more than a paragraph. Because the constraint is invisible from the portfolio side, it can only enter the room if somebody writes it down. A liquidity requirement announces itself when a payment is due. A concentration limit announces itself when a holding drifts. A tax constraint announces itself once a year, long after the decisions that set its size were taken, and by then the decisions are not reversible.

Why does the constraint belong to the holder and not to the holdings?

Because of what it is measured against. Everything else in a governing document is measured against the portfolio: a 5 per cent cap on a single holding is measured against Rs 500 crore, a 50 to 70 per cent equity band is measured against the same Rs 500 crore, and a minimum credit standing is measured against the fixed income sleeve. All three bases sit inside the document. The base for a tax constraint is a fact about the entity signing the document, and that fact was settled before anybody chose a single holding.

Take the Anantara Multi-Asset Portfolio seriously here rather than treating it as a label. The holder is a charitable endowment. A reader who works with private wealth may be about to conclude that none of this applies to them, and a reader at an institution may be about to conclude that the whole question is settled by the institution's status. Both conclusions are wrong in the same way. The mechanism is identical for a private holder and for an institution. The only thing that changes between them is the holder's own position, and that position is a fact to be confirmed rather than a fact to be assumed.

An exempt holderA holder whose position means a realised gain is not reduced. Whether any particular holder is in that position is settled by the tax authority and that holder's own advisers, never by a general description. is simply the case where the reduction happens to be nothing. Exemption is not a different mechanism but the same mechanism at one end of its range. A mandate written on the assumption of one position and then run under another has not made a small error. The mandate has run the wrong arithmetic on every trading decision for the whole period.

When does a tax constraint actually bite?

At realisationThe moment a position is actually sold and a gain or loss stops being a valuation and becomes a completed transaction., not at appreciation. A holding that has risen and has not been sold has produced a number on a valuation statement and nothing else. The moment it is sold, the gain stops being an opinion about price and becomes a completed event with a counterparty, a date and a settled amount.

Most people picture this the wrong way round, and the picture matters more than the vocabulary. The instinct is that holding something valuable creates the constraint, in the way that holding a house creates a rates bill. It does not. A decision to sell is also a decision to realise. Realisation attaches the tax constraint to trading activity rather than to holding, and holding alone is where most people first imagine it sits.

Feel it in the household version. A grandmother holds gold jewellery that has risen many times over since it was bought. Nothing happens, year after year, no matter how much the price moves. The event that matters is the morning she takes it to be sold, and until that morning arrives the appreciation is a story about price rather than a transaction anyone has to account for. She could hold it for another twenty years and the sequence would never start.

Two things follow at once, and both are construction facts rather than accounting facts. The first is that a portfolio can carry very large unrealised gains and a very small constraint, or the reverse, and the size of the portfolio does not say which. The second is that the constraint is generated by the same activity that generates trading costs. The constraint therefore belongs in the same conversation as turnover, not in a separate appendix at the back.

How Tax Constraints Affect Portfolio Construction

There are three routes by which this reaches construction, and none of them requires knowing a rate. All three are mechanical, and all three change what a document should say rather than what a market will do.

Route one is arithmetic: the reduction comes out of a return the holder never sees in full, so a stated return and a kept return are two different quantities carrying the same digits. The Anantara Multi-Asset Portfolio returned 14.2 per cent gross over its stated twelve months. The endowment finished with that figure less whatever the reduction was, and the record does not contain the second term.

Route two is about timing. Once the constraint attaches at the moment of sale, the timing of a sale stops being a by-product of a portfolio decision and becomes a decision in its own right. Somebody is choosing it whether or not they know they are choosing it. A mandate that never mentions realisation has not avoided the decision. The mandate has delegated the decision, silently, to whoever writes the trade tickets.

Route three is the one that catches experienced drafters, and it is why the tax constraint belongs with the drafting of documents rather than with the keeping of accounts. A rebalancing rule that trades more will realise more. So the rebalancing paragraph and the tax paragraph are connected by a quantity that neither of them mentions, and a committee tightening a band for perfectly good discipline reasons has also, in the same motion, enlarged the tax constraint on the whole mandate.

The right-hand column above is worth as much as the left. A tax constraint does not change the equity band, does not change the 5 per cent limit on a single holding, does not change the minimum credit standing on the fixed income sleeve, and does not change why any particular holding was wanted in the first place. A tax constraint changes what a given amount of trading costs the holder. The claim is narrower than the one people usually make for it.

What exactly is the wedge, and who controls each term?

The relationship follows, and it is short. The reduction, called the tax wedgeThe gap between what a portfolio returned before any reduction and what the holder finished with after it. The wedge is a difference between two returns, not a payment schedule., is a rate applied to the realised part of a return. Two terms, multiplied. There is nothing else in it.

Which of the two terms sits where is the point. The rate comes from outside the document entirely and no committee can change it. The realised shareThe part of a period's return that came from positions actually sold, as opposed to the part that is still showing on a valuation. is produced by how the portfolio is run, so the document decides it whether it means to or not. One term is given from outside; the other is written by the document, every time a trading rule is written.

Now the honest bit. Using that relationship on the Anantara Multi-Asset Portfolio requires the realised share as a number, and the invented record does not contain one. The record contains turnover, a different quantity. Turnover measures traded value, not the share of the return that was crystallised. The 34 per cent turnover figure therefore serves as an explicitly stated stand-in for the realised share, named as such at every appearance.

The stand-in is a device for teaching rather than a measurement. A real mandate would take its realised figure from its own settlement records rather than from a turnover ratio. Turnover and realisation move together, which is why the stand-in is useful, but they are not the same number.

With the stand-in declared, the arithmetic runs in one line. The Anantara Multi-Asset Portfolio returned 14.2 per cent gross over the stated twelve months. Thirty-four per cent of 14.2 is 4.828, carried as 4.83 percentage points. Those 4.83 points are the realised part of the return on this stand-in, and every later step multiplies them.

The missing term matters. One of the two terms is in hand and the other is not, so no wedge can be produced. The whole line of wedges the relationship allows can be produced instead, one for each rate somebody might supply, and the line teaches more than a single answer would.

Run the line, then. At a rate of one tenth the wedge on a 4.83 point realised share is 0.48 points. At two tenths it is 0.97 points. At three tenths it is 1.45 points. Not one of those three rates is in force anywhere. Each is an input chosen to show that the wedge rises in a straight line with the rate rather than in steps.

Why is turnover the term the mandate controls?

Here a tax constraint stops being an accounting matter and becomes a construction matter. The realised share is the only term in the wedge that anybody in the room can change. A tax constraint therefore changes what a rebalancing rule costs, and with it how the rule gets written.

Watch it move. Halve the turnover on the Anantara Multi-Asset Portfolio from 34 per cent to 17 per cent. On Rs 500 crore that is Rs 85 crore of traded value rather than Rs 170 crore. On the same stated stand-in, 0.17 times 14.2 is 2.414. The realised part of the 14.2 per cent gross return falls from 4.83 points to 2.41 points. Every wedge on the line above halves with it, at every rate, and the rate never has to be known.

Be careful about what the halving does and does not license. Trading less is not therefore better. Trading less changes other things too: it changes how far the actual weights drift from the policy weights of 60, 30 and 10 before anyone pulls them back, and drift is the thing the rebalancing paragraph existed to control in the first place. Two constraints pull in opposite directions, and writing them in one place lets somebody see the pull.

What is the wedge worth, set beside something already established?

A wedge of 0.97 points means nothing on its own. Is that a lot? Compared with what? A wedge only becomes legible the moment it is set beside the part of the return that came from something other than market exposure. The mandate was arguably run to produce exactly that part.

The split is built under the committee memo and is borrowed here rather than rebuilt. Over the stated twelve months the Anantara Multi-Asset Portfolio returned 14.2 per cent gross against the composite benchmark's 12.6 per cent, a gross excess of 1.6 percentage points. With the risk-free rate at 6.5 per cent, the benchmark stood 6.1 points above it, and at a beta of 1.08 the exposure alone accounts for 1.08 times 6.1, or 6.588 points, on top of 6.5, giving 13.088. Subtract that from 14.2 and 1.112 points is left over, so about 0.49 of the gross excess was carrying more of the market and about 1.11 points was not.

The crossing is the sharpest step available. If the wedge is a rate times 4.83 points, the rate at which the wedge equals 1.11 points is 1.11 divided by 4.83. The division gives about 23 per cent. Checked forwards, 0.23 times 4.828 is 1.11044, and that rounds to 1.11.

At a rate somewhere near 23 per cent, on this stated stand-in for the realised share, the wedge would be the same size as the entire part of the gross outperformance that the extra market exposure did not explain. The 23 per cent is a solved crossing point and not a rate in force anywhere. The endowment's own position is not stated anywhere in its record.

The crossing is easy to overstate, and the overstatement is tempting. Be careful with what it licenses. The crossing does not mean the manager achieved nothing. The narrower claim is that a pre-tax returnA return measured before any reduction for the holder's position. A pre-tax return is what a portfolio produced, not what the holder finished with. comparison and an after-tax returnA return measured after the reduction the holder's position causes. An after-tax return is the figure a holder can actually spend or reinvest. comparison can rank the same year differently, and that a mandate reporting only the first has not yet told the holder what they kept.

What must a mandate say about tax, and what must it never say?

Everything above lands in one paragraph of a governing document, and the paragraph is short. The paragraph carries three things and refuses a fourth.

The paragraph names the holder's position as a fact to be confirmed rather than a fact assumed in the drafting room. It names who confirms that position, the holder's own advisers together with the published material, and it names when the confirmation is refreshed. The third thing is an instruction on realisation, usually to record and report it rather than to optimise anything. And it never states a rate or a treatment. A document carrying a rate becomes wrong on the day the treatment changes, and nobody in the room finds out until the year is over.

There is a fifth line worth adding, and it is what all of this establishes: a sentence recording that the rebalancing paragraph and the tax paragraph were read together, and naming the realised share as the quantity that joins them. The sentence costs nothing to write, and it is the only thing standing between a committee and a constraint nobody chose.

How does a committee actually use this, on a Tuesday?

Rukmini Deshpande's committee cannot compute a wedge. Two of the three inputs a result would need are not on its own record. The committee does something smaller and more useful instead, and it fits on one sheet of a quarterly pack.

The first line is the confirmation. Somebody states, in writing, that the endowment's position has been confirmed with its advisers within a named interval, and that the confirmation is on file. The position itself is not stated on the portfolio record. The portfolio record is the wrong place for it, and the manager is not the person who determines it.

The second line is the realised figure. Not turnover as a proxy, the stand-in used above, but the actual realised amount taken from settlement records for the period. The moment that figure sits on the pack beside the 34 per cent turnover, the committee can see whether its stand-in was even close. Nobody can run that check while the realised figure is missing.

The third line is the join. One sentence recording that when the rebalancing band was last set, the realised amount it would produce was looked at. One sentence is the difference between a constraint the committee chose and a constraint that arrived.

An analyst looking in from outside runs the same three questions in reverse. Is the reported return gross or after any reduction? If gross, does the pack anywhere say what the reduction was? And if the answer is no, then the 1.6 point gross excess over the composite benchmark is a portfolio fact rather than a holder fact, and the two are being quietly treated as one.

The household version fits on the back of an envelope, and it is the same three lines. A person selling part of a long-held holding to fund a wedding writes down what they will actually receive after everything, not what the statement says the holding is worth. Then they write down who told them the treatment, and when. Then they check whether the sale they are about to make was chosen for the wedding or chosen by a rule they set years ago and forgot about. Same discipline, same order, no rate needed for any of it.