How large should an emergency fund actually be?

An emergency fund is the least interesting part of a financial plan and the part that most often determines whether the plan survives. It earns a modest return, appears on no leaderboard, and does nothing visible for years at a time. Then a job ends, a parent is hospitalised or a roof fails, and it quietly becomes the reason a decade of investing is not liquidated at the worst possible moment.

The standard advice — six months of expenses — is a serviceable default that hides three questions worth answering properly: six months of which expenses, why six, and where should the money sit.

Which expenses actually count

The fund is not sized against your normal spending. It is sized against the spending that continues even when income stops. Under real pressure, discretionary spending falls sharply and immediately — nobody funds holidays and restaurant meals out of an emergency fund. What does not fall is the fixed portion, and that is what you are insuring.

Investments are the other common error, in the opposite direction. SIPs should be excluded, because pausing them is exactly the right response to a genuine emergency. Including them inflates the target and delays the point at which you are actually protected.

How many months you need

The number of months is a function of how long it would realistically take to replace your income, and how many things could go wrong at once. Six months is the default because it approximates a typical white-collar job search with a margin for the notice-to-offer gap. Move up or down from there based on your own situation.

A worked example: essentials of ₹60,000 a month — ₹28,000 of home loan EMI, ₹18,000 of groceries and utilities, ₹8,000 of school fees, ₹6,000 of premiums and transport. A single-earner household should hold around ₹3.6 lakh. At ₹20,000 a month set aside, that takes eighteen months to build, which is why this is a project rather than a decision.

Build it in stages, not in one go

An eighteen-month project with no visible progress is easy to abandon. Breaking it into rungs gives you a sequence of genuine wins, and each rung removes a specific category of risk.

1. 1

   One month — stop the bleeding

   Enough to cover a single cycle of essentials. This is the difference between an unexpected bill and a credit card balance at 40% a year.

2. 2

   Three months — absorb a real shock

   A medical event, a redundancy with notice pay, or a major repair becomes manageable rather than destabilising. Most of the benefit of an emergency fund is captured here.

3. 3

   Six months — the default target

   Full cover for a normal job search. At this point you can decline a bad offer, which is a financial return that never shows up in a return calculation.

4. 4

   Nine to twelve months — for concentrated risk

   Worth reaching if you are the only earner, your income is variable, or your role is specialised. Beyond twelve months, extra cash is usually better deployed elsewhere.

Where to keep it

There is exactly one requirement: the money must be available on the day you need it, in full, without a loss. Every other consideration — return, tax efficiency, elegance — is subordinate to that. In practice a split works better than a single instrument.

• One month in a savings account. Instant, no process, no market hours. This is the layer you actually touch.
• Two to three months in a sweep or short-tenure fixed deposit. Ideally broken into several smaller deposits so you can withdraw ₹50,000 without breaking ₹3 lakh. A fixed deposit is predictable and, for most people, entirely adequate for this layer.
• The remainder in a liquid or overnight fund. Redemption typically lands the next working day, with somewhat better returns than a savings account and very little price movement.

What does not belong here: equity of any kind, ELSS, gold, real estate, or anything with a lock-in. The reason is not that these are bad assets — it is that emergencies correlate with bad markets. Recessions produce job losses and falling share prices at the same time, so the moment you need the money is disproportionately likely to be the moment it is worth least.

An emergency fund invested in equity is not an emergency fund. It is an investment you have promised to sell at the worst possible time.

What holding it really costs

The objection to a large buffer is always opportunity cost, so it is worth pricing rather than hand-waving. Take the ₹3.6 lakh fund above, held for five years at around 6.5% instead of invested at 12%.

So the buffer costs roughly ₹1.4 lakh over five years, or about ₹2,350 a month. Set against what it prevents — redeeming long-term investments in a downturn, borrowing at 40% on a card, accepting the first job offer that appears — that is inexpensive. It is also not a fee, since the money is still yours and still growing, just more slowly.

The comparison also flatters equity by assuming it delivers 12% over exactly the five years you happened to hold it. The whole reason this money is held in cash is that it cannot depend on that assumption being true on any particular date.

After you use it

Using the fund is not a failure — it is the fund working. The failure is not rebuilding it. Treat replenishment as a temporary, high-priority SIP: pause discretionary investing, direct the surplus back into the buffer, and restore at least the three-month rung before resuming normal contributions. Emergencies are not evenly spaced, and the period right after one is not a period of reduced risk.

Two final maintenance points. Revisit the target whenever your fixed costs change materially — a new loan, a new dependant, a move to a more expensive city — because a fund sized against your old outgo quietly covers fewer months than you think. And keep it in an account you do not use for daily spending, since the practical enemy of an emergency fund is not a market crash but gradual, unremarkable erosion.