Where Should an NRI Hold an Emergency Fund: The Currency-Matching Rule, and Why a Remitted India FD Is the Wrong Place for True Emergencies

A reader in Toronto lost his contract job in late 2025 with about six weeks of runway before rent and his car loan would start biting. His "emergency fund" was a Rs 18 lakh NRE fixed deposit in India earning 7.1%, tax-free, which he was rightly proud of. The problem revealed itself the week he needed it. Breaking the FD eleven months in meant forfeiting almost all the interest, the repatriation back to Canada took four working days through his bank's SWIFT process, and the rupee had weakened against the Canadian dollar since he remitted, so he converted at a worse rate than he bought in at. By the time the money landed in his Canadian account, he had already put two months of expenses on a credit card at 21%. He had saved diligently and still got caught, because he held the right amount of money in the wrong currency, in the wrong country, behind the wrong amount of friction.

The 30-second answer: Hold your emergency fund in the currency of your liabilities, which for almost every NRI is the host country, not India. Your rent, mortgage, medical excess and a sudden flight home are priced in USD, GBP, AED or CAD, so the fund must clear in hours with no exchange-rate risk. A host-country high-yield savings or money-market account pays roughly 4 to 5% in the US and UK and 2.75% in Canada in 2026, with same-day access. A remitted India FD pays more on paper, 6.5 to 7.5%, but sits behind a 2 to 5 day SWIFT round trip and a currency conversion at the worst moment. Hold 3 to 6 months of essential expenses, 6 to 12 if your visa is employer-tied, and keep a smaller NRE liquid balance only for genuine India-side obligations.

This guide assumes you already know the difference between NRE, NRO and FCNR accounts and how repatriation works; if not, start with the accounts guide. What follows is the part that actually goes wrong: the single rule that decides where the money lives, the maths that shows why a higher Indian rate loses to a lower host-country rate for this specific job, how to split a fund across two countries when your life genuinely straddles both, and the tax and liquidity trade-offs that the FD-versus-savings debate usually gets backwards.

The one rule that settles the location debate: match the currency of what you owe

An emergency fund is not an investment. Its single job is to be there, in full, in the currency you will spend it in, on the day a shock arrives. Every other property, yield, tax treatment, repatriability, is secondary to that. The mistake NRIs make is treating the emergency fund like the rest of the India corpus, optimising it for return, when the correct optimisation is for certainty of access.

That gives you the rule the entire decision hangs on: hold the fund in the currency of your liabilities. Write down what the money would actually pay for in a real emergency. A job loss means rent or a mortgage instalment, utilities, groceries, your car payment, insurance premiums, all denominated in your host-country currency and debited from your host-country account. A medical emergency means a hospital excess or an out-of-pocket bill in USD, GBP, AED or CAD. Even the one expense that points back to India, a last-minute flight home for a family illness, is bought with a host-country card. Almost nothing in a genuine emergency is paid in rupees from India.

So the currency of your liabilities is the host country, and that is where the fund belongs. This is not a patriotism question or a "where do you trust the banking system more" question. It is an asset-liability matching question, the same discipline a bank uses when it funds rupee loans with rupee deposits. If your liabilities are in pounds and your emergency fund is in rupees, you have taken on currency risk on the one pot of money that exists specifically to remove risk. The rupee has trended weaker against the dollar, pound and dirham over almost every multi-year window, which means the structural bias works against you, but the deeper point is that you should not be exposed to the move in either direction on emergency money. You are insuring against a shock, not running a currency view.

There is a narrow, honest exception, and it is worth stating plainly because it is real for some readers. If your liabilities genuinely sit in India, you are weeks from returning permanently, you support parents whose bills you pay directly, or you are carrying an Indian home loan you service from India, then to that extent your emergency fund should sit in India in rupees. The rule does not say "always host country." It says "match the currency you owe in," and for most working NRIs that currency is overwhelmingly the host one, with a thin rupee tail.

Why a host-country savings account beats a higher-yielding India FD for this job

The instinctive objection is the rate. In 2026 a good NRE fixed deposit pays in the region of 6.5 to 7.5% (SBI sits around 6.25 to 6.60% for one to ten years, while Axis runs 6.75 to 7.25% and Yes Bank up to roughly 7.75%), and the interest is tax-free in India and fully repatriable. A US high-yield savings account pays up to about 5.00% APY, a UK easy-access account around 4.5 to 5.0%, a UAE high-yield account 4 to 6.25% on the right balance tier, and a Canadian everyday HISA like EQ Bank about 2.75%. On a naive comparison the India FD wins on three of four host markets. So why is it the wrong place?

Because for emergency money the rate is close to irrelevant, and the two things that actually matter, speed of access and freedom from conversion risk, both favour the host-country account decisively.

Start with speed. A host-country high-yield savings or money-market account is same-day or next-day money. You move it to your spending account with an app transfer and pay your landlord that afternoon. A remitted India FD is a multi-step chain: break the deposit, move the proceeds to your NRE savings account, raise a repatriation, wait the 2 to 5 working days a SWIFT transfer typically takes to clear, and only then convert and land it abroad. When the emergency is "rent is due Friday and I just lost my job," a four-day pipeline is not an emergency fund, it is a slow fund.

Then the conversion. Bringing money back from India means selling rupees for your host currency at whatever the rate is on the day you are forced to act, plus the bank's spread, which on retail conversions is routinely 1 to 2% away from the interbank rate. You do not get to wait for a good rate in an emergency; you take the rate you are given. So the tax-free 7% you earned in rupees can be partly or wholly eaten by an adverse move and a wide spread on the way out, exactly as it was for the Toronto reader.

Put real numbers on the comparison so the gap is concrete. Say you need a GBP 12,000 emergency cushion as a UK-based NRI, and you are choosing between a UK easy-access account at 4.6% and remitting the equivalent to an NRE FD at 7.0%.

In the UK account, GBP 12,000 at 4.6% earns about GBP 552 in a year, sitting in cash you can spend the same day. In the India FD, you would remit roughly GBP 12,000, say at 105 rupees to the pound, giving about Rs 12,60,000. At 7.0% tax-free that earns Rs 88,200 in a full year, which converted back at the same rate is about GBP 840. On the surface the FD is GBP 288 ahead.

Now the counterfactual that matters, because emergencies do not respect the one-year holding period. Suppose you need the money at month seven. The FD broken before maturity does not pay 7.0%; it pays the rate that prevailed at booking for a seven-month tenure, often nearer 5.0 to 5.5%, minus a penalty of around 0.5 to 1%, so call it an effective 4.5% for the seven months held, roughly Rs 33,000 of interest, about GBP 314 gross. Against that, the rupee has slipped 3% over those seven months, costing you about GBP 360 on the principal when you convert back, and the bank's 1.5% conversion spread costs another GBP 180. Your GBP 840 of fantasy interest has become a real loss of roughly GBP 226 versus where you started, before you even count the four days you spent on a credit card waiting for the wire. The UK account over the same seven months simply earned about GBP 320 and was available the morning you asked for it.

That is the whole argument in one calculation. The India FD wins a race that an emergency fund never runs (held untouched for a full year or more), and loses badly at the race it actually runs (broken early, converted under pressure). The higher headline rate is compensation for illiquidity and currency risk, which are precisely the two things an emergency fund cannot afford to take on.

Within the host country, prefer a genuine high-yield savings account or a money-market account at an established, deposit-insured institution over a chequing account paying near zero. The national average US savings rate is about 0.38%, so the difference between a default big-bank account and a 4 to 5% online account is real money on a six-month cushion, and you give up nothing in access to get it. In the UK, an easy-access account or a cash ISA (up to GBP 20,000 a year of tax-free interest) does the same job; Premium Bonds at a 3.6% prize rate are a worse fit because the "return" is a lottery and not every pound earns. In the UAE, accounts like Mashreq's NEO PLUS Saver reach 6.25% on a salary-transfer condition, which is unusually generous for instant-access cash and worth using if you qualify. In Canada, EQ Bank's roughly 2.75% everyday rate beats the big banks' near-zero savings rates while keeping CDIC protection.

Where an NRE liquid balance still earns its place

None of this means you keep zero rupee liquidity. It means you size the India-side balance to genuine India-side needs rather than treating it as your primary emergency fund. There is a real, recurring set of obligations that are best met from India in rupees, and holding a modest buffer there is sensible asset-liability matching, not a contradiction of the rule.

The India-side liquidity bucket covers things like your parents' regular support or a sudden medical bill for them, the maintenance and property tax on a flat you own in India, an India home-loan EMI if you service it locally, school or college fees for a child studying in India, and the routine cost of a planned trip home. For these, money already in India in rupees is the right tool, and moving it from a host-country account would mean paying to convert in the wrong direction.

Hold this in an NRE savings account for instant access, topped up by a short NRE fixed deposit ladder if the balance is large enough to bother. The reason to favour NRE over NRO for this buffer is that NRE balances and their interest are fully repatriable and the interest is tax-free in India, whereas NRO interest is taxable in India with TDS, so for fresh money you are sending from abroad, NRE is cleaner. The trade-off is that NRE savings rates are modest, HDFC pays about 2.50% on NRE savings, so do not park a large sum idle there; keep only what you would plausibly need on short notice in the savings account and ladder the rest into one-year NRE FDs you can break individually if needed. If you want the deepest comparison of where to put the medium-term rupee money, read NRE FD versus FCNR FD, because the FCNR route lets you hold the deposit in foreign currency in India and sidesteps rupee risk on that slice.

One subtlety that catches people: deposit insurance in India, through the DICGC, covers only Rs 5,00,000 per depositor per bank, principal and interest combined, aggregated across all branches of that bank. If your India-side liquidity plus FDs at a single bank exceed that, you are uninsured above the limit at that bank. For an emergency-adjacent buffer this matters more than for long-term money you are actively monitoring, so if your rupee buffer is large, spread it across two banks or lean on the most stable institutions. Host-country deposit insurance is generally more generous (USD 250,000 under FDIC, GBP 85,000 under FSCS, CAD 100,000 under CDIC), which is one more quiet point in favour of holding the bulk of the fund abroad.

How many months to hold, and the one cost residents never face

The standard advice of three to six months of expenses is a fine starting point, but it was written for someone whose job loss does not threaten their right to live in the country. For an NRI, it sometimes does, and that changes the sizing.

Anchor the number on essential expenses, not your full lifestyle: rent or mortgage, utilities, groceries, insurance premiums, minimum debt repayments, and basic transport. Strip out discretionary spending, because in a real emergency you cut it anyway, and an emergency fund sized on your comfortable monthly burn is larger than it needs to be and drags on your investing. For a salaried NRI on a stable footing, three to six months of that essential figure is right.

Push toward six to twelve months in three situations. First, if your visa is employer-tied, an H-1B in the US, a Skilled Worker visa in the UK, a standard UAE employment visa, a closed work permit in Canada, then losing your job can start a clock on your legal right to remain. In the US, an H-1B grace period is up to 60 days to find a new sponsor or change status; a UAE employment visa typically gives a grace period after cancellation; in those windows you may be job-hunting and visa-fixing at once, and a thin fund forces a bad decision. Hold more. Second, if you are a single earner supporting a family or dependents on a linked visa, because the household has no second income to fall back on. Third, if you support parents in India out of current income, since those obligations do not pause when your salary does, and they are part of your essential burn even though they sit in another country.

Then add the line residents never have to budget for. An NRI emergency is unusually likely to involve a sudden return to India, a parent in hospital, a bereavement, a family crisis, and a last-minute international flight for one person, let alone a family, can run Rs 80,000 to Rs 2,00,000 booked days before travel, sometimes more in peak season. Build a discrete flight-home line into the fund rather than assuming your general cushion absorbs it, because this expense tends to arrive at the same time as the emotional and logistical chaos that makes everything else harder. Holding it in host-country cash is correct, because you buy that ticket with a host-country card the night you find out.

A worked split across host country and India

Here is how the pieces fit together for a concrete reader, so you can adapt the structure rather than the exact numbers. Take Priya, a UK-based NRI on a Skilled Worker visa, single earner, with essential monthly expenses of GBP 2,400, who sends her parents in Pune Rs 40,000 a month and owns a small flat there with annual property tax and maintenance of about Rs 60,000.

Because her visa is employer-tied and she is a single earner, she targets the upper end, nine months of essentials. Nine times GBP 2,400 is GBP 21,600, which she holds in a UK easy-access account at 4.6%. On top of that she adds a GBP 1,500 flight-home line in the same account, for a host-country emergency pot of GBP 23,100, all same-day money in the currency she actually spends. At 4.6% that pot earns roughly GBP 1,063 a year sitting completely liquid.

For the India side, she sizes the rupee buffer to genuine India obligations, not to her whole life. Three months of parental support is Rs 1,20,000, plus the Rs 60,000 annual flat costs, plus a small medical buffer for her parents of Rs 1,00,000, totalling about Rs 2,80,000. She holds Rs 1,00,000 in an NRE savings account for instant access and ladders the remaining Rs 1,80,000 into three NRE FDs of Rs 60,000 each maturing at four-month intervals, so something is always close to maturing and breakable without forfeiting much. Crucially she keeps the single-bank total under the Rs 5,00,000 DICGC limit, so the whole India buffer is insured.

Now the counterfactual that shows the discipline paying off. Suppose Priya had instead done what the Toronto reader did and held the entire cushion, say the GBP equivalent of about Rs 26 lakh, in one big NRE FD at 7.0% because the rate looked best. If she then lost her job and needed roughly GBP 14,000 over four months, she would break the FD early (forfeiting most of the 7%, netting maybe 4.5% for the months held), wait several days for the wire, and convert about Rs 14.7 lakh back to pounds under whatever rate and spread applied that week. A 2% adverse rate move plus a 1.5% spread on that conversion is about GBP 490 of pure friction cost, on top of credit-card interest while she waited. The split structure costs her that GBP 490 in avoided friction and hands her the host-country money the same afternoon, while the India FDs keep quietly funding her parents in rupees with no conversion at all. The split is not the higher-yielding choice on a spreadsheet; it is the one that actually works on the worst week of the year.

Edge cases

You are within a year or two of returning to India permanently. As your liabilities shift back to rupees, so should the fund. In the run-up to a permanent return, deliberately rebuild the emergency cushion in India in rupees, because that is where your rent, bills and life will be priced. The currency-matching rule does the work in both directions; it just usually points abroad while you are still earning abroad.

You hold an FCNR deposit and call it your emergency fund. An FCNR FD is held in foreign currency in India, so it removes rupee risk, which is a genuine plus over an NRE rupee FD for emergency money. But it is still a fixed deposit in India behind the same repatriation pipeline, and FCNR deposits usually carry a minimum tenure (commonly one year) with no interest paid on a very early break. It is better than an NRE rupee FD for this purpose because of the currency match, but it is still slower and stickier than a host-country savings account, so use it for medium-term foreign-currency savings rather than as your front-line cash.

Your host-country tax treats savings interest unkindly. In the US, savings and money-market interest is ordinary income taxed at your marginal federal rate plus state tax, which can clip a 5% account meaningfully; in the UK, a Personal Savings Allowance shelters the first GBP 1,000 of interest for a basic-rate taxpayer (GBP 500 for higher-rate) and a cash ISA shelters more; the UAE has no personal income tax, so the headline rate is the real rate; Canada taxes interest as ordinary income. None of this changes the location rule, because the tax cost of holding emergency cash where you can reach it is small relative to the cost of being unable to reach it. But within the host country, route the cushion through the most tax-efficient liquid wrapper available, a cash ISA in the UK is the obvious one, before reaching for a higher nominal rate elsewhere.

You are tempted to put the fund in liquid mutual funds or money-market funds for a better yield. This is reasonable for the upper slice of a large fund, and a host-country money-market fund is close enough to cash to qualify for part of the cushion. An Indian liquid mutual fund is not, for the same reason the FD is not: redemption proceeds land in an Indian account in rupees behind the repatriation and conversion pipeline, and for an NRI from the US or Canada there is also the PFIC reporting overhead on Indian funds. Keep the genuine emergency layer in cash or near-cash in the host country, and let liquid funds sit one layer out, as the bridge between emergency cash and long-term investing rather than as the emergency fund itself. How that bridge fits the whole picture is covered in NRI portfolio asset allocation.

The closing read

The honest read is that NRIs lose money on emergency funds not by saving too little but by holding the right amount in the wrong currency and the wrong country, chasing a tax-free 7% that evaporates the moment they actually need the cash. The rate is a distraction. An emergency fund has exactly one job, to be fully there in the currency you spend, on the day you ask, and a host-country high-yield savings or money-market account does that job while an India FD does not.

So for most working NRIs the recommendation is firm: hold three to six months of essential expenses, six to twelve if your visa is employer-tied or you are a single earner, in a high-yield savings or money-market account in your host country, in host-country currency, at a deposit-insured institution, plus a discrete line for a last-minute flight home. Keep only a smaller NRE balance sized to genuine India-side obligations, parents, an Indian property, local fees, laddered into short NRE FDs and kept under the Rs 5 lakh DICGC limit per bank. The exception, and only this exception, is the reader whose liabilities genuinely sit in India: someone weeks from a permanent return, or servicing an Indian loan from India. For everyone else, the FD-versus-savings instinct gets it backwards. Optimise the emergency fund for certainty, optimise everything else for return, and never let the two pots blur into one.

Related guides

• NRE, NRO and FCNR accounts: which one for what
• NRE FD versus FCNR FD: the currency-risk choice
• Building an NRI portfolio: asset allocation across two countries
• Your first month abroad: setting up your money
• Currency hedging for NRI investors
• All Banking guides
• All Investments guides
• All Jobs guides

This guide is educational and general in nature. It is not individual financial advice. Interest rates, deposit-insurance limits and visa grace periods change, and the right split for you depends on your currency exposure, visa status, dependents and country of residence, so confirm your specific position with a qualified adviser before committing a large emergency fund to any single account or country.