Monthly budget planner: start from what is spoken for

Take a household with ₹1,25,000 of take-home pay, a rented flat, a car loan, a child in school and a monthly transfer to parents. Asked for a budget, they write eleven rows totalling ₹83,800, 67.0 percent of take-home, and conclude that ₹41,200 a month is free.

Then they total three months of statements and card bills. Month one: ₹1,20,747, because the quarter's school fee of ₹21,000 fell in it. Month two: ₹1,35,397, the month the family health premium of ₹26,400 renewed. It spent ₹10,397 more than came in, and the difference went on a card. Month three, with nothing lumpy but a ₹1,200 annual subscription: ₹1,00,193. The three-month average is ₹1,18,779, which is ₹34,979 more than the remembered budget, or 41.7 percent.

None of the three months was unusual. Every month has something, a different something each time, so memory files each one as a one-off.

Layer one is fixed and contractual: rent, EMIs, monthly premiums, SIPs and other mandates, AutoPay bills, and transfers you have promised someone. Most leave by NACH or UPI AutoPay before you see the salary. Here that is nine rows and ₹64,347, 51.5 percent of take-home. Without the ₹10,000 SIP, which is saving rather than spending, it is ₹54,347, or 43.5 percent.

Layer two is irregular but certain: bills that come once a quarter or once a year, as predictable as rent, just less often. Look back twelve months, not three, because a three-month window catches only the lumps that happened to fall in it. This household has six: the school fee at ₹21,000 a quarter, ₹12,000 of annual school charges, the ₹26,400 health premium, ₹13,800 of car insurance, a festival month that runs ₹36,000 over normal, and ₹2,400 of annual subscriptions. That is ₹1,74,600 a year, or ₹14,550 a month, 11.6 percent of take-home.

Layer three is variable: the three-month average of everything else. Eight rows, ₹38,232, 30.6 percent. Use the average, not the round number you would like it to be.

Layer four is what is left: ₹7,871, 6.3 percent of take-home. Not ₹41,200. It is small, but ₹10,000 of saving already sits in layer one. The household is not failing; it was planning against a number that never existed.

A sinking fund turns layer two into a monthly transfer to a separate account, so the lump is paid from money set aside for it, not from the month it lands in. The divisor is the months until the bill, not twelve.

The family health premium of ₹26,400, started a full year ahead, is ₹2,200 a month. Started five months before renewal it is ₹5,280. Started two months before, ₹13,200, which is more than the household's whole layer four and is how the premium ends up on a card. The festival month works the same way: ₹36,000 is ₹3,000 a month from a year out and ₹9,000 a month from four months out. A quarterly school fee never gets a long runway: ₹21,000 a quarter is ₹7,000 a month.

Four layers, filled in order, each from a different source. Layer one comes from the last statement: every debit repeating at the same amount. Layer two comes from twelve months of statements and card bills, plus anything you know is coming that history cannot show, such as a wedding next year. Layer three is the average of three months. Layer four is the subtraction.

The worked column is the illustrative household. Replace it with your own and leave the order alone.

Pick a fixed date a few days after salary: fifteen minutes, three parts.

Five minutes on layer one. Did anything new start repeating: an EMI, a mandate from a forgotten trial, a raised premium? A ₹2,097 AutoPay line is ₹25,164 a year, and mandates charge until someone cancels them.

Five minutes on layer two. What lump falls due in the next three months, and is it already set aside? If the health premium renews in five months and nothing is set aside, the number is ₹5,280 a month from now.

Five minutes on layer three. Roll the average forward: drop the oldest month, add the newest. Say groceries came to ₹13,900 in month four, ₹2,224 or 19.0 percent over the ₹11,676 average. Past a 15 percent trigger, ask one question: was this a one-off, such as guests, or a new level? If a level, the new three-month average of ₹12,563 is the forecast, and layer four shrinks by the difference.

For next month's forecast alone, without a full budget, see the page on predicting next month's bills.

The household above makes about 126 payments a month, most of them small UPI payments through GPay, PhonePe or Paytm. Typing them into a spreadsheet at 20 seconds each is 42 minutes a month and 504 a year, and that is before anyone reads a number. So the first job of an app is to import real outflows from records the bank already produces, across every route, including NACH and transfers, which a payment app's own history never shows.

The second is flagging the irregular items: a yearly lump is invisible in a three-month view and obvious in a twelve-month one. The third is predicting next month: layer one plus whatever from layer two falls due. That is a calculation, not a judgement.

A spreadsheet is enough when the month is simple: few payments, a stable layer one, and a person who will do the review. It also beats any app at the future history cannot show: a wedding next year, a school change. They belong in layer two before they happen.

A budget allocates money you have. It cannot create money, and it cannot outrun interest. Three cases where it is the wrong first move.

An income shortfall. Suppose the same household's take-home falls to ₹95,000. Layers one and two now take ₹78,897, 83.0 percent, leaving ₹16,103 for a variable layer that averages ₹38,232: ₹22,129 short every month. No grocery limit closes that. The levers are all in layer one, the rent, the EMI, the SIP that can be paused, and in income. Trimming layer three alone moves the shortfall onto a card more slowly.

Debt at card rates. An ₹80,000 card balance carried at an illustrative 3.5 percent a month, 42 percent a year, plus 18 percent GST on the interest, costs ₹3,304 a month. That is 42.0 percent of the household's ₹7,871 layer four, and ₹39,648 a year if the balance merely stays where it is. Clear it before the finer points of budgeting and before new long-term investing; no allocation of the rest beats a 42 percent cost. Keep the sinking fund for the next premium running, though, or the premium goes straight back on the card.

No problem to solve. If layer four is large and stable and saving happens by mandate on salary day, a detailed budget adds effort without changing a decision. Review layers one and two, and stop.