Investing outside India

What it adds, in two items

Buying a foreign asset from India adds exactly two things. It adds ownership of businesses that never listed here, in industries that may have little listed presence on an Indian exchange. And it adds a currency leg, because you buy in rupees and the asset lives in another currency. You will one day want rupees back.

Everything else in the subject is a consequence of those two. The routes differ in how the second one is handled and when. The tax differs because the wrapper is classified by what it holds. The costs differ because someone has to convert the money and someone has to keep the shares. None of that is exotic once the two additions are separated, and most of the confusion comes from treating the pair as one thing called “international exposure.”

The second addition deserves its own sentence, because it is the one that arrives unannounced. An unhedged foreign holding carries a currency position sized at the entire holding. Not a sleeve of it, not a hedged fraction — all of it. If someone offered you a standing bet on the rupee equal to the full value of a fund, you would at least ask what it cost. It comes bundled, priced at nothing explicit, and most fact sheets will not draw your attention to it.

The arithmetic of that leg is short and it belongs elsewhere. A rupee investor earns the foreign asset's return in its own currency, multiplied by the currency's move against the rupee. On illustrative round numbers chosen so the multiplication is easy to redo, a foreign index up 10% with the dollar up 4% against the rupee gives an Indian unhedged holder 14.4%; the same index up 10% with the dollar down 4% against the rupee gives 5.6%. Same fund, same year, 8.8 percentage points apart, and the fund manager did nothing differently. The rupee article works that table through in full, including what hedging costs and why the hedged holder is not simply giving up a tailwind.

This article takes it from there: what the routes do to that leg, what the wrappers cost, why the tax is not the tax you are used to, and — the part usually skipped entirely — what the diversification argument is assuming when it says foreign assets help.

A listing venue is not an economy

Start with the claim doing the most work in most coverage: that adding foreign equities gives you exposure to another economy. That is a claim about where the money is earned. What you actually bought is defined by where the shares trade. Those are different facts, and nothing in the label tells you how far apart they are.

An Indian software exporter is listed in Mumbai and bills its customers in dollars. Its earnings rise and fall with corporate technology budgets in North America and Europe, and with the rupee. A holder of an Indian index fund already owns that revenue stream. It is counted as domestic because of the venue, not because of the economics.

The same trade runs the other way. Large companies listed abroad sell into India, and some of them derive a meaningful part of their revenue here. Buying a global index does not exclude India; it re-buys a slice of it through a foreign listing, at that foreign listing's valuation, with a currency leg attached.

The country label measures the venue, and the venue is a legal and historical accident — where the company chose to raise capital, decades ago, under the rules of the time. Nothing about the label promises that the revenue is domestic.

Two consequences follow, and they point in opposite directions, which is why this is worth sitting with rather than reading past.

• The economic diversification added by a foreign holding is smaller than the listing labels suggest, because both portfolios already contain revenue from the other side. You are not moving from zero foreign exposure to some foreign exposure. You are moving along a line whose starting point you never measured.
• The genuinely new element is larger than most people notice, because the currency leg is not diluted at all. Whatever the overlap in revenue, the exchange rate applies to the full holding. The one part of the trade that is unambiguously new is the part nobody was shopping for.

There is a structural point underneath, and it survives every change in index composition. An index of one country's listed companies can only contain what has listed in that country. If an industry has no significant listed presence in India at a given moment, no Indian index can hold it, however important the industry is to the Indian economy or to the reader's own working life. This article names no sector, because listings change and any list of absences goes stale. The question generalises better than any answer to it: instead of asking what an index holds, ask what it could not hold.

Which turns the usual framing around. The reason to look outward is not that foreign markets go up; nobody knows that, and this article forecasts nothing. It is that a portfolio built from one country's listings inherits that country's listing history — its sector mix, its ownership structures, its disclosure regime — whether or not anyone chose it. That is a fact about the wrapper, not a prediction about returns.

The routes, and what actually differs between them

Four ways to hold foreign assets from India, plus one that arrives without being chosen. They look interchangeable from a distance. They differ on who converts the money, when the value is struck, what the price you pay is anchored to, and what paperwork attaches to you personally. One of them sends rupees out of the country, which a resident individual may do up to an annual ceiling set per person — the Liberalised Remittance Scheme, the remittance scheme below.

Three of those five are Indian mutual fund schemes, and the difference between them is not cosmetic. The first buys shares. The second buys a fund that buys shares. The third is bought and sold by you on an exchange rather than issued and redeemed by the fund house. That last distinction decides what price you get, which is the subject of the next section.

One further route exists and this article does not describe it — buying foreign securities through the international financial centre at GIFT City. Its permissions, mechanics and tax treatment are a separate subject, and a half-described route is worse than an unmentioned one.

The paperwork asymmetry is the practical divider. Units of an Indian scheme are an Indian asset: you hold rupee-denominated units issued in India, whatever the scheme holds underneath. Shares held directly abroad are a foreign asset in your name, with a foreign-country tax authority upstream of your dividend and a disclosure obligation downstream in your return. Whether that trade-off is worth it depends on what you want — direct holding lets you own a specific company, which no Indian scheme can do for you.

What the wrapper does to the price you pay

Here is the specific mistake this section exists to name. An investor buys an Indian-listed feeder ETF on a foreign index during Indian market hours, on a day when the foreign market is shut, and pays more for the units than the holdings behind them are worth — without any error message, any warning, or anything on the screen looking unusual.

The mechanism is worth having, because it applies to every exchange-traded wrapper on anything. An ETF's market price is held close to the value of its holdings by an arbitrage: large intermediaries can create new units by delivering the underlying and redeem units by taking the underlying back, so a price above the value of the holdings invites creation and a price below invites redemption. That valve is what makes the price track the value. Neither the fund house nor the exchange enforces the link.

Shut that valve and the price has nothing anchoring it. If a scheme cannot issue fresh units — for any reason, and there are several — the only thing setting the screen price is what Indian buyers will pay Indian sellers for a fixed pool of units. A wave of demand then moves the price of the wrapper rather than the size of the fund. Nothing about the foreign shares has changed. You are paying a premium for access to a queue.

The time-zone version is milder and permanent. When you trade an Indian-listed feeder on a foreign index at 14:00 IST, the foreign market that sets the value of the holdings is closed. The screen price is a live opinion about a stale value. Some drift is honest price discovery — the world has moved since the foreign close, and buyers are pricing that in. Some of it is simply the imbalance of who happened to want units that afternoon. From outside, the two are not distinguishable.

Call it the shut-market premium. It is a property of the wrapper rather than a view about the asset, and the thing that exposes it is unglamorous: the fund publishes its own value alongside the screen price, so the gap between the two is observable before anyone transacts. A wide gap is information about the wrapper, not a signal about the underlying market. The gap can run in the seller's favour as readily as against the buyer's, which is exactly why it is not a cost anyone budgets for.

The other two fund routes have no exchange price and no premium, and pay for that with a different property. You transact at a NAV struck later, on a calendar that is not yours. A fund of funds sits behind an offshore fund that strikes its own value on its own trading days, so the Indian NAV mechanically reflects a value from a different clock; the scheme's own cut-off — generally 3:00 pm IST for a non-liquid scheme — then decides which day's value you get. This is not tracking failure and it is not a fee. It is a consequence of two markets keeping different hours, and it is why comparing an Indian scheme's day-to-day movement against a foreign index's day-to-day movement is comparing two different days. How a NAV is struck and when covers the general machinery.

Why the tax is not the tax you are used to

The capital-gains rules an Indian investor is likeliest to know are the ones that apply to a domestic equity fund. Those rules are not a rule about equities. They are a rule about a defined category, and an internationally-invested scheme is usually not in it.

The machinery is worth learning, because it survives rate changes. A scheme is equity-oriented when it puts more than 65% of its total proceeds into domestic equity shares. Clear that test and the familiar treatment follows: gains on units held beyond 12 months take the long-term rate of 12.5% above an annual threshold of ₹1.25 lakh, and gains realised sooner take 20%.

The load-bearing word is domestic. A scheme that holds foreign equities does not satisfy the equity-oriented test by holding equities — the test names the place, not the asset class. Nor is such a scheme automatically in the other named bucket: a Specified Mutual Fund is one holding more than 65% in debt and money-market instruments, which a fund holding foreign shares plainly is not. Failing both named tests, it falls into the ordinary capital-gains treatment that applies to assets with no special category of their own.

This article states no rate and no holding period for that third bucket, because that would be a figure without a verified source, and a tax figure copied forward for three years is worse than no figure at all. The general framework, and where to look up which bucket applies, is in the capital gains article. What matters here is the shape of the reasoning: the classification is arithmetic on what the portfolio actually holds, and it is stated in the scheme document. It cannot be read off the scheme's name, its benchmark or its category label.

Direct foreign holdings sit in a different regime again, and add three obligations that units of an Indian scheme do not. Tax is collected at source when money leaves the country under the remittance scheme. A foreign company's dividend is generally taxed in the paying country before it reaches you, with any relief depending on a treaty and on a credit claimed at your end. And holding a foreign asset directly attaches a disclosure obligation in your own return that holding an Indian scheme does not.

None of those three is stated here with a rate, a threshold or a form, and that is deliberate. They are named so that the reader knows the questions exist before choosing the direct route, not after. The mistake worth naming is not paying too much tax. It is discovering an annual obligation in the year someone finally asks about it.

A fourth question sits further out and this article does not answer it either. Shares held in your own name in another country may bring that country's estate or inheritance tax into the picture on death, on a basis that depends on the country and on how the holding is structured — a question that attaches to direct holdings and not to units of an Indian scheme, which are an Indian asset whatever they hold underneath. It is raised here because it is the obligation least likely to be discovered by the person who could still do something about it.

The diversification argument, and what it assumes

The standard case for foreign assets is that two markets which do not move together produce a portfolio calmer than either. That is arithmetic and it is correct as arithmetic. It is also a model, and a model is a set of assumptions. Naming them is not a caveat appended to the argument — it is most of what an investor needs from the argument.

Five assumptions sit underneath, and every one is known to be imperfect.

• Stable correlations — the benefit is computed from how two markets moved together over some past window. Correlation is a property of a sample, not of a relationship. It moves, and there is a mechanism for the direction it tends to move in under stress: a global withdrawal of risk appetite sells many markets at once from a single decision, so a shared cause overrides whatever made them differ. The diversification is smallest on the days it is most wanted. That much is mechanism. How far it goes in any given episode is an empirical question with a different answer each time, which is exactly why a single correlation number is a poor summary of it. How to read a correlation window covers what the number can and cannot tell you.
• Near-normal returns — the mathematics measures risk as variance, which treats a quiet 1% day and a violent 8% one as the same kind of event at different scales. Return records repeatedly show joint tails — days when several apparently unrelated things fall together — fatter than a pairwise correlation implies, and how much fatter is itself estimated from a past sample and differs by market and period. A model calibrated on ordinary days understates what happens on the extraordinary ones, and understates it exactly where the loss is concentrated.
• A single-period horizon — the framework asks what happens over one period, then the answer gets applied to someone saving for twenty-five years and spending in rupees. Those are not the same problem. Currency volatility over a month and currency drift over decades are different objects, and a model that treats risk as one number has silently collapsed them.
• Two independent legs — the rupee return multiplies the asset return by the currency move, and the tidy arithmetic treats those as separate. They are not always separate. In a global risk-off episode the dollar and foreign equities can move together in ways that make the currency leg a cushion or an amplifier depending on the regime, and which one you get is not knowable in advance. The dollar's channels are traced separately.
• Market-cap weights as neutral — a global index weighted by market value is often presented as the default, unopinionated choice. It is a decision: to own whatever the world has listed, in proportion to what the market currently prices it at. That embeds a view that prices are the right weights, and it means the portfolio's country mix is set by other people's listing and valuation history.

None of that says the argument is wrong. It says the argument is conditional, and the conditions are the interesting part. A reader who has absorbed the mechanism can update when a correlation regime changes. A reader who has memorised a conclusion cannot.

Keep two kinds of claim apart while reading anything on this subject. That two markets driven by different domestic conditions can offset each other is a mechanism claim, and it follows from how the arithmetic works. That international exposure has improved outcomes for Indian investors is an empirical claim about a period, and it is contested, sensitive to the start and end dates chosen, and different for every currency pair and every window. Anyone quoting the second should be quoting a period with it. Rolling returns exist precisely because a single window's answer is mostly a fact about the window.

Which is why this article states no allocation and no range. The size question has no general answer, and the honest version of it is not “how much international exposure is right” but a list: what your existing holdings already earn abroad, which of your future costs are set in a currency you do not earn, and how much of a currency swing you can watch without acting on it. The second of those is worked through in the rupee article, and the general framework for sizing anything against a goal is in asset allocation.

What it costs, listed honestly

Costs on this route come in two kinds: the ones quoted to you and the ones that arrive as a difference between two prices. The second kind is larger and less discussed.

The trade-off, stated plainly and without resolving it: the fund routes cost an ongoing percentage and remove almost all the administration; the direct route costs conversion spread, foreign withholding and personal paperwork, and in exchange lets you own a specific company rather than an index. Neither is free, and the cheaper one on paper is not always the cheaper one after the second kind of cost is counted.

One more item that is not a cost but behaves like one. A foreign holding drifts against the rest of the portfolio for two reasons at once — the asset moved and the currency moved — so it wanders from its intended size faster than a domestic holding does. Whatever you do about that, rebalancing a two-legged holding is a different exercise from rebalancing a one-legged one, and the same is true of any asset priced in a foreign currency, including gold.

Where to look at this

Nearly everything above is checkable rather than arguable, and the checking is arithmetic on published series. FNOTrader's Mutual Funds app runs on the full AMFI NAV history — around 34 million rows, updated nightly at 22:30 IST — and every scheme's NAV series in it is already in rupees, which is the currency an Indian holder actually experiences.

For an internationally-invested scheme that matters more than usual. It simulates both a lumpsum and a fixed monthly instalment — a systematic investment plan, or SIP — on any scheme and period, and reports XIRR — the internal rate of return for cashflows landing on irregular dates — alongside invested versus value and maximum drawdown, plus rolling-return distributions across every available window rather than one trailing figure. The two-legged return is what those numbers are computed on, because the NAV already contains both legs.

What no tool settles is the classification question in the tax section or the hedging position in the section above it. Both are terms of the specific scheme, and both are in its scheme document.

Common questions

Continue reading

More in Advanced Investing · App: Mutual Funds · Definitions: glossary · Free tools: calculators · All: every article