Your Food budget is four budgets: splitting groceries, khaja, eating out and delivery in Nepal

A cousin sent me a screenshot last month with one line circled: Food & Drinks, Rs 32,000, Bhadra. Family of four, Kathmandu, two salaries. "Is this normal?" she asked. There was no way to answer. That one number held the Bhatbhateni run, the Kalimati vegetables, two office khaja habits, a birthday dinner and a run of rainy-evening delivery orders. Four very different decisions, added into one line that could not tell her which one to change.

It turns out Nepal's own inflation statisticians reached the same conclusion, and recently.

What NRB changed, and why it matters to your budget

For a decade, Nepal's CPI treated a plate of momo at a restaurant and a kilo of rice from the kirana pasal as the same kind of spending. Both sat in "food and beverage", which carried 43.91% of the old 2014/15 basket.

The rebased index, built for the first time from the Nepal Living Standards Survey IV, splits them. Following the UN's COICOP 2018 standard, restaurants, alcohol and tobacco now live under non-food and services. The weights, from NRB's technical note on the 2023/24 base year:

NRB notes that if the three moved groups were folded back in, food would stand at 45.79%. So the share of spending that is food-ish did not shrink. It got split, because cooking at home and eating out behave differently: different prices, different drivers, different people making the call.

Part of that jump from 2.92% to 8.00% is likely a measurement change rather than pure behaviour. NLSS-IV was the first round to record meals eaten outside the home separately for every household member, instead of one household-level estimate. Asking each person found more eating out. That is the point of this post in one sentence: look at eating out separately and you find more of it than you expected.

Khaja, not lunch, is the big line

The restaurant group is itself broken into items in the index. This is the part worth reading twice:

Afternoon khaja and chiya together weigh 4.03, just over half of everything Nepal spends in restaurants and accommodation (calculation: 4.03 ÷ 8.00). The sit-down lunch, the thing people picture when they say "eating out", is under a third. The goods list NRB prices includes buff, chicken and veg momo, chowmein, samosa and buff, chicken and mutton khaja sets.

Nationally, khaja is the largest single eating-out item in the basket. It is also the one nobody budgets for, because nobody decides to spend Rs 2,750 a month on it. They decide, twenty-two separate workdays, to spend Rs 125.

How big the whole Food line should be

Before splitting, it helps to know whether the total is in a sane range. NLSS-IV reports the share of spending that goes to food (meals out included) by income band. Kathmandu Valley households spend a smaller share on food than anywhere else in the country:

The first two columns are from the NLSS-IV summary report (Tables 2 and 3). The last column is my calculation: spending × food share ÷ 12. The figures are 2022/23 prices, and the same grocery basket costs noticeably more in 2083, so read them as a floor.

On those numbers, a middle-band family of four would spend around Rs 26,800 a month on food at 2022/23 prices (Rs 6,700 × 4). My cousin's Rs 32,000 is not alarming for a two-salary household three years of inflation later. The total was never the problem. The mix was invisible.

Splitting one line into four

Split by where the decision gets made, not by what was eaten. A momo bought at the office canteen and a momo delivered on a Friday night are the same food and completely different decisions.

A fifth line, Festival and guests, earns its place in Asoj and Kartik, when a single week of Dashain hosting can outweigh a month of groceries. The Dashain family cost breakdown shows how fast that line fills; the festival sinking fund is how to pay for it without the credit card.

Stop at four or five. Every extra sub-category is one more choice at the moment of logging, and a tracker only works if logging takes two seconds.

A worked example: the Rs 32,000 Bhadra

Here is my cousin's month, split. The categories are real; the per-item amounts are illustrative round numbers, so substitute your own.

Seen split, the month tells a different story. Groceries are 61% of the line and hard to move much. Khaja is Rs 5,500 a month, which is Rs 66,000 a year, more than the two sit-down meals and more than the delivery orders. It was the only line nobody had noticed. The daily Rs 200 leak is the same pattern at the level of a whole day; here it sits hidden inside a category that looked like "food".

Delivery deserves its own row for a separate reason: it carries costs a restaurant visit does not. Foodmandu's published delivery charge is Rs 50 for the first 2 km and Rs 25 for each km after. A 4 km order is Rs 100 in delivery alone, so five orders is Rs 500 a month and Rs 6,000 a year in fees. That money buys no food. Filed under one Food line, it is invisible.

The receipt check that belongs in Eating out

When a sit-down meal does go in the Eating out line, check the bill against the menu. For years, a Rs 1,000 menu price turned into Rs 1,243 at the till: a 10% service charge, then 13% VAT on top of both, which is 24.3% more than the printed price.

After an August 2022 Kathmandu District Court ruling, the Department of Commerce, Supplies and Consumer Protection required menu prices to include VAT and service charge, and fined restaurants Rs 2 lakh each for billing above them. If you see those two lines added on top today, the menu price was not the real price. Your tracker entry should be the amount you actually paid, not the menu figure.

What the split changes

A split category does three things one Food line cannot.

It shows which habit moved. Your total rising Rs 4,000 in a month could be Dashain guests, a price rise in oil and meat, or a new office with a better canteen. Only the sub-category tells you, and each has a different response.

It makes the budget honest. A Food cap of Rs 30,000 invites "I'll cook more next week" as the fix every time. A Rs 4,000 cap on Khaja and chiya is a decision you can keep on a Tuesday afternoon. If you use the 50/30/20 split, groceries sit in the needs half and khaja and delivery in the wants 30. One Food line smudges the boundary that matters most.

It gives the monthly review something to look at. The 15-minute monthly review asks what changed. "Food went up" is not an answer. "Delivery doubled in the rains" is.

What you actually need to know

• One Food line is several budgets. NRB's own index now separates eating out (8.00%) from food bought to cook (35.49%), and afternoon khaja is the biggest eating-out item in it.
• Split by decision, not by dish. Groceries, khaja and chiya, eating out and delivery each respond to a different fix. Four or five sub-categories is the ceiling.
• The small daily line is usually the surprise. Two people's office khaja at Rs 125 a day is Rs 66,000 a year, typically more than the restaurant meals people feel guilty about.

Questions, or a food split that works better for your household? Write to parjanya57@gmail.com.

This post is part of the Nepal Money Basics guide, in the understand-your-money section.