Emergency Fund: What a Buffer Covers, Measured in Months
The months of cover division converts a buffer into a length of time. Put in what actually leaves the household in an ordinary month, put in what can be spent inside the week a shock allows, and it divides the second by the first. The answer comes back in months. A shock does not pause a yearly bill, so the division asks for everything that leaves rather than only what is committed.
Most households that run this arithmetic honestly get an answer under one month, and that is a description of a position rather than a mark against anybody. The invented household worked through below lands under one month too. Its second income halved for five months in the year being measured, through nothing anybody decided, and what survived is what survived. No line is drawn anywhere in this arithmetic, so no reader falls below one. The division computes a result and then stops.
The arithmetic itself is one division and nothing else. All of the difficulty sits in the two numbers fed into it, and both of them are routinely entered too generously, in the same direction, by people acting in complete good faith. The top number invites a household to count money it cannot actually spend this week. The bottom number invites it to leave out the bills that only arrive once a year. The two mistakes compound rather than cancel, and a household making both will read an answer close to eleven times the true one. Every note under every input in this calculator exists to stop that, rather than to make a single division look rigorous.
What does this working tool compute?
The division computes a length of time. Not an amount, not a score and not a rating. Two rupee figures go in and a number of months comes back, and that change of unit is the whole reason for running it. A household that knows it has set aside Rs 31,320/- knows a fact about a balance. Time is the thing a shock takes away, and money is only the form in which time is stored. A household that knows what that Rs 31,320/- covers knows something it can plan around.
The bufferMoney that can be reached and spent inside the time a shock allows. A buffer is defined by how fast it moves rather than by how safe or how large it is. goes on top. The money that leaves in an ordinary month goes underneath. The result is called months of coverThe buffer divided by what leaves in a month. Months of cover answers how long a household could go on paying for its ordinary life with nothing coming in., and it is worth reading out loud, because the sentence it produces is blunt in a way the rupee figure never is. Rs 31,320/- sounds like an achievement, and it is one. The same amount stated as a number of weeks does not sound like anything except what it is.
Think about the stall in the market lane for a moment. Ashok Bhosale runs a tailoring counter there, and if it shut tomorrow the rent on the room, the electricity, the cooking gas and the school terms would go on arriving exactly as before. Nobody posts a letter pausing them. The only question that matters on the day the counter shuts is how many of those arrivals the household can meet out of what it already holds, and that question has a numeric answer. The months of cover division is the arithmetic that produces that answer, and all of the difficulty lies in feeding it correctly.
What goes into the top line, and how many documents does it take?
The bottom number is the harder of the two and the one people guess, so it comes first. The figure required is what actually left the household in a month with nothing unusual in it, including a share of everything that only arrives once or twice a year. For the Bhosale household at 31 March of year two that figure is Rs 42,770/- a month.
No single document carries that number. That is exactly why it is the input most often estimated, and why the estimate is almost always low. It is assembled from three places. The last three bank statements give the ordinary month: rent, groceries, electricity, cooking gas, mobile, broadband, fuel, medicines and the society maintenance. A yearly calendar of the items that do not arrive monthly gives the rest: the school terms for Ira Bhosale, the insurance premiums, the vehicle papers, the one festival month that is never an ordinary month. The card and loan statements give whatever a debt demanded that month. The first, a twelfth of the second and the third, added together, are the figure.
| Where the bottom number is built from | What is read off it | For this invented household |
|---|---|---|
| The last three bank statements | Every payment that repeats monthly, taken from the statement lines rather than from memory | Rs 34,770/- a month |
| A written list of the once-a-year and twice-a-year items | The annual total of school terms, premiums, vehicle papers and the festival month, divided by twelve | Rs 8,000/- a month |
| The card statement and any loan schedule | What a debt demanded in the month being measured, taken from the printed line | Rs 0/- in March |
| What actually leaves in an ordinary month | The figure that goes underneath the division | Rs 42,770/- |
The third row is worth a sentence. In March the two-wheeler loan had already closed, its thirtieth instalment paid in January, so nothing was demanded by it. Until January the same row carried Rs 3,150/- a month and the household's money out ran at Rs 45,920/-. The measurement therefore uses Rs 42,770/- rather than Rs 45,920/-. The second figure is history, and what leaves the household now is the first.
How many documents does the bottom input need before it can be filled in honestly?
What counts as being in the buffer?
Now the top of the division, where the second mistake lives. The question is not the household's net worth. The buffer question is much narrower, and it is about speed.
The reachability testThe single question that decides whether a holding counts: can it be turned into spending inside the week a shock allows. Value and safety are not part of it. is one line long. Can this money be spent inside the week a shock allows? Not is it real. Not is it safe. Not is it large. Can it be spent this week, in a shop, for groceries, without asking anybody for anything. A holding either passes that test or it does not, and the ones that fail are not disqualified for being poor holdings. The failure is one of speed alone.
The Bhosale household holds Rs 3,67,887/- in all. How much of it passes the reachability test is the next figure.
Run the test across the household's holdings and the ladder sorts itself into three groups by speed rather than by size. The account the salary lands in holds Rs 10,567/-, spendable today. The account kept aside and not touched once in the year holds Rs 31,320/-, also spendable today. Together they are Rs 41,887/-, and that is the whole of what passes. The recurring deposit stands at Rs 64,000/- of deposits paid in, real and slower. The provident fund at Rs 84,000/- is not close. Gold at the household's own estimate of Rs 1,40,000/- and the two-wheeler at Rs 38,000/- are worth what somebody will pay on the day. That day is not a week from now, and the price is not a figure anybody knows in advance.
The bar carries the sentence this guide is built around. The Bhosale household is worth Rs 2,96,293/- once what it owes is taken off, and it can reach Rs 41,887/- inside a week. The reachable amount is about a ninth of what it holds and 14.1 per cent of what it is worth. Net worth and resilience are two measurements of the same household on the same day, and they disagree violently. Neither is wrong. The two measurements answer different questions, and only one of them is the question a shock asks.
The household holds Rs 64,000/- in a recurring deposit. Does that amount go into the buffer input?
Why is a deposit maturing in eight months left out?
Because the buffer input is a question about the week and a deposit is an answer about a date. A week and a date are different things, and the arithmetic breaks quietly when they are mixed.
Breaking a depositEnding a deposit before the date it was booked for. Breaking a deposit usually takes some days to arrange and usually costs something, and both are set by the bank that holds it. before its date is possible and households do it. Breaking one is not instant, involves a form and a branch or an application, and usually costs something. The cost at any particular bank is that bank's own arrangement, and it moves. Where the Rs 64,000/- matters to a particular reading, the question goes to the bank that holds the deposit, and the framework those arrangements sit inside is published by the Reserve Bank of India at rbi.org.in.
The tool does not throw the deposit away; it moves it to a second line, so a household reads a fast number and a slower number instead of one blended number that is true of neither. For the Bhosale household the fast line is Rs 41,887/- and the line that includes the deposit ended early is Rs 1,05,887/-. Both are real. The two lines just arrive at different speeds, and a shock that lands on a Tuesday cares a great deal about which one it can reach.
Which outgoings figure belongs underneath the division?
Two figures are available and both are true descriptions of this household. The choice between them is the one most often made without anybody noticing they made it.
Committed outgoingsWhat cannot be stopped this month: rent, groceries, power, transport and any payment a debt demands. The items that arrive once a year are outside it by definition. at 31 March are Rs 34,770/- a month. The committed figure is built to answer a different question, namely what has to be met this month and cannot be postponed. By construction it leaves out anything that does not arrive monthly. What actually leavesEvery rupee out over an ordinary month, including a twelfth of the items that only arrive once or twice a year. is Rs 42,770/- a month. The figure is the same Rs 34,770/- plus Rs 8,000/- a month of school terms, premiums, vehicle papers and the festival month.
The gap is Rs 8,000/- a month and the argument for closing it is simple. A shock does not pause the school terms, and it does not pause an insurance premium, and a household in its fifth week without income does not get to tell the school that this is a difficult period. The yearly items arrive on their own schedule whatever has happened to the income. Any denominator that excludes them is measuring how long a household can meet part of its life, and calling the answer months of cover.
There is a second reason and it is about direction. The committed figure is smaller, so it always produces a longer answer. Not sometimes longer. Always longer, for every household, for every buffer, without exception. The same numerator over a smaller denominatorThe figure the buffer is divided by. The denominator decides the answer as completely as the buffer does, and it is chosen rather than looked up. is always a bigger quotient. An input that can only ever err in the comfortable direction is worth being suspicious of on that basis alone, before any question of whether it is right.
Which outgoings figure belongs underneath the division for this household, Rs 34,770/- or Rs 42,770/-?
What does the division do to the two numbers?
The division turns rupees into months, and it does so in a straight line. The straight line sounds like a technicality and is the most useful property the arithmetic has.
Because months of cover is buffer divided by outgoings, and the outgoings figure sits still while the buffer is what changes, the relationship between the two is a straight line through the origin. A rupee added gains a fixed fraction of a month. Rs 42,770/- added gains exactly one month. Added again, it gains another. For this household one month of cover has a price, and the price is Rs 42,770/-, the same figure as its monthly outgoings. The match is not a coincidence.
The price of a month converts an abstract intention into a number that can be worked with. A household that puts aside Rs 3,500/- a month is buying roughly a twelfth of a month of cover every month. About a year of that buys one month. The price is neither a discouraging fact nor an encouraging one. It is an exchange rate, and knowing an exchange rate is strictly better than not knowing it.
How much added to the buffer buys this household exactly one more month of cover?
Commit to an answer before the panel below reveals it. Rs 31,320/- in the account kept aside, against Rs 42,770/- a month leaving. How long does it cover?
Move what the buffer holds and watch both denominators answer at once.
One thing moves here: what the buffer holds, from nil to Rs 2,00,000/-. Both outgoings figures are held at the Bhosale household's own, Rs 42,770/- for everything that leaves and Rs 34,770/- for the committed figure, so neither has to be chosen blind. The panel opens at Rs 31,320/-, exactly what sits in the account kept aside at 31 March. The strip underneath the bars matters too: the three markers are this household's own reachability positions, and they fill in as the control passes them.
Two movements are worth noting. The buffer input reads speed and not wealth, so with the control at nil the household still holds Rs 3,26,000/- of other things. Moving it slowly right widens the gap between the two bars: the two denominators do not disagree by a fixed amount, they disagree by a fixed proportion, so the further right the control goes the more the choice of denominator is worth.
How should the output be read, as a description rather than a score?
The default reading is 0.73 months. Rs 31,320/- divided by Rs 42,770/- a month comes to about three weeks. Add the Rs 10,567/- sitting in the account the salary lands in, on the ground that it is spendable the same day, and the reading becomes Rs 41,887/- over Rs 42,770/-, or 0.98 months. Include the recurring deposit ended early and the reading is Rs 1,05,887/- over the same denominator, or 2.48 months, arriving days later than the first two.
Read against committed outgoings instead, the account kept aside gives 0.90 months rather than 0.73. The difference is about five days, and both readings stand side by side rather than one being quietly chosen. 0.73 and 0.90 are the same household on the same evening. A reading describes a position on a date, so it should always be written down with both of its inputs attached.
Now the larger movement, and it is not the denominator at all. Changing the denominator moved this reading by about five days. Changing what counts as buffer moves it by weeks. From the account kept aside alone to both accounts is 0.73 to 0.98, and adding the deposit ended early takes it to 2.48. How much money counts as reachable moves the answer several times harder than which outgoings figure sits underneath. The reachability test is the larger of the two questions, and the choice of denominator the smaller.
Something else was going on across the same year, and it explains something the arithmetic cannot. Across year two the account kept aside was not touched once, held about Rs 30,000/- and was credited Rs 1,140/- of interest. Across the same year the card was charged Rs 6,447/-. The household was paying its contracted 3.5 per cent a month on a card balance while holding money doing a great deal less, and nobody decided that. The groceries went on the card in September because the salary account was empty on the day, and the buffer sat in a different account with a different purpose attached to it in somebody's head. Richard Thaler's work on mental accounting describes exactly this: money is not treated as one pool but as several, each with a label. Mental accounting is a mechanism, not a lapse, and it is one reason the reading and the position can drift apart.
What a household does with the number, and what somebody on the other side of a counter does with the same arithmetic
A household uses it as a stock take with a deadline attached. The valuable part is rarely the quotient; it is the two lists that had to exist before the quotient could. Writing down what actually leaves in a month, from three documents rather than from memory, is often the first time that total has been seen at all, and sorting holdings by how fast they move is often the first time anybody has asked which of them is money on Thursday. For the Bhosale household the exercise produced Rs 42,770/- and Rs 41,887/-, and the 0.98 that follows is almost an afterthought. Somebody on the other side of a counter runs the same arithmetic for a different reason: a bank's own hardship desk, a school office deciding whether to accept fees in parts, or a lender assessing a request to reschedule, all of them are asking a version of how long can this household meet this without help. Each organisation then applies its own policy to the answer. The same division serves a household counting its own weeks and a stranger assessing a request, and in neither use does the number become a verdict on anybody.
Where is every number found, document by document?
The useful part of a working tool is knowing which sheet of paper each figure is copied from. Every figure below has a document and a line. A field noteThe short line under an input naming the document a figure is read off. A field note says where to look, not what the figure means. names where to look and nothing else. Interpretation is a separate job from location.
Only one input in this calculator can be read off a single document, and it is the smaller of the two. The buffer is a set of balances on the day of the reading, one line per account, and there is nothing to construct. The outgoings figure has no home document at all and must be built. It therefore carries three notes instead of one, and it is the figure a careful household writes down and keeps rather than recalculating from memory every time.
| Input | Where the figure is found | Left to a separate question |
|---|---|---|
| What actually leaves in a month | The last three bank statements, plus a twelfth of a written yearly list, plus any amount a card or loan demanded in that month | Which of those lines could be reduced, which is a different exercise entirely |
| What the buffer holds | The balance of every account that can be spent from today, read on the day of the reading | Where that money should sit, a separate question about products |
| The slower line, shown separately | The deposit passbook or certificate, at the amount paid in | What ending it before its date costs, which is set by the bank holding it |
| The output | Computed, not found. Buffer divided by outgoings, stated in months | Whether the answer is enough, which no arithmetic can settle |
The failure: two generous inputs, both pointing the same way
A household sits down with the same tool and enters everything it holds against its committed outgoings. Neither entry is dishonest and neither is careless. Rs 3,67,887/- really is what the household holds, and Rs 34,770/- really is what it must meet this month. The tool divides them and returns 10.58 months.
The true reading on the definitions used here is 0.98 months. The gap is not a difference of opinion about method. It is the distance between a household that believes it has most of a year and a household that has three weeks, and the tool cannot detect either input.
Worked in two steps, the fall shows that neither mistake alone explains it. Correct only the denominator, so Rs 3,67,887/- over Rs 42,770/-, and the reading is 8.60 months, still wildly long. Correct only the numerator, so Rs 41,887/- over Rs 34,770/-, and the reading is 1.20 months, still flattering by about a week. Both corrections are needed to arrive at 0.98, and that is precisely why the two mistakes are dangerous together. The errors do not offset. They multiply.
The provident fund is not spendable on Thursday. The gold is worth what somebody pays on the day and the household would be selling it into its own emergency. The two-wheeler is how Meghna Bhosale gets to work, so selling it to survive a shock costs the household the income it is trying to protect. And the Rs 8,000/- a month of school terms and premiums arrives on its own calendar whether or not there is a salary that month.
A household enters everything it holds, Rs 3,67,887/-, against its committed outgoings of Rs 34,770/-. What does the tool return?
What does this tool refuse to do?
No number of months is named as a target, and none is marked on any scale. Every axis is a ruler with no flag planted in it.
The reason is that the right length for any particular household depends on things no tool can see. How steady the income is, and whether it is one income or two. Whether a second earner could pick up work inside a month. Who else in the household could help and who is depending on it. How the household's health stands. Whether the work is the kind that stops entirely or the kind that thins out. Whether there is cover for the specific shock in question. Cover is a matter for the insurer, and it sits under the framework published by the Insurance Regulatory and Development Authority of India at irdai.gov.in. A tool that printed a number of months would be answering a question it has none of the information for, and it would be answering it identically for every reader.
There is a second reason and it is about what a printed number does to a reader. A mark on a scale stops being a description the moment somebody stands next to it. The household reading 0.73 months would not learn anything from a flag at some other figure; it would simply be told it had failed, by a calculator that knows nothing about it. The Bhosale household's buffer is thin because its second income halved for five months. A halved income is not a character finding, and no arithmetic should be arranged to look like one.
Why does this tool mark no number of months on any of its scales?
What sits outside this arithmetic, and why no figure for it appears here
The division itself is universal and needs no jurisdiction. Two things around it are not. The charge for ending a deposit before its date, and the time it takes, are set by the bank holding it and sit inside the framework published by the Reserve Bank of India at rbi.org.in. Whether a particular shock is covered by insurance, and what any policy does, sits with the insurer and inside the framework published by the Insurance Regulatory and Development Authority of India at irdai.gov.in. Cover and a buffer are different things doing different jobs.
What is worth writing down once the panel is closed?
Three things, and none of them is a figure to aim at.
The first is the two lists rather than the quotient. The first list is what actually leaves, built from three documents. The second is what can be spent inside the week, sorted by speed. Both lists are durable. The quotient falls out of them in a second and can be recomputed any evening.
The second is the label that has to travel with any reading. A reading is written as 0.73 months on committed outgoings and against the account kept aside only, or as 0.98 months on everything that leaves and against both accounts, never as the bare number. A reading without both of its definitions attached is a number somebody can argue about forever, including the household that wrote it, six months later.
The third is the exchange rate. For this household one month of cover costs Rs 42,770/-, and that single fact turns every future decision about setting money aside from a vague intention into an arithmetic one. It also makes the position legible rather than shameful: 0.73 months is not a failing grade, it is Rs 31,320/- measured against a life that costs Rs 42,770/- a month to run, and both of those numbers came from documents rather than from anybody's opinion.
References
| Source | Document | Where |
|---|---|---|
| Reserve Bank of India | Material on deposit accounts and term deposits in India, named because what happens when a deposit is ended before its date sits inside that framework and with the bank holding it, and no figure for it is stated here | rbi.org.in |
| Reserve Bank of India | Customer conduct and fair practices material on banking in India, named for the existence of that framework and for the fact that account terms sit inside it | rbi.org.in |
| Insurance Regulatory and Development Authority of India | Material on insurance in India, named only because cover is the thing a buffer does not replace, and no premium, limit or policy figure is stated here | irdai.gov.in |
| Richard Thaler | The published work on mental accounting, named because the observation that money is treated differently depending on which account it sits in belongs to it | named in the text rather than quoted |
The Bhosale household, Meghna Bhosale, Ashok Bhosale, Ira Bhosale and Sahyadri Freight Services Private Limited are invented.
Educational material. Not advice on any investment, tax, budget or market position.