The price of a basket of life.

The blue line is one continuous headline on the 2024 base: no seam, no splice. It exists because the NSO did the hard part itself: it back-cast the new index to 2013, re-reading the whole last decade on the new ruler, and it prints live levels from 2025. Joining the two (both base 2024=100) gives a single 2014–2026 series. The one gap, the 2025 year-on-year, is computed at the handoff, from the back-cast's 2024 levels to the live 2025 levels. No 2012 figure enters this line.

The line lightens to the left of January 2025. That paler portion is the NSO's back-cast: a reconstruction of history on the 2024 basket, not a direct measurement, and the NSO may restate it. From January 2025 the stroke turns dark: live, collected prices. The change of weight is the change of evidence. Read the dark part as fact and the lighter part as the NSO's best estimate of what the new ruler would have shown.

The faint grey line is the retired 2012 ruler, drawn for comparison. Where the two overlap they read the same months slightly differently: a different basket at work, not a contradiction. This is why the two bases are never joined into one number: they are different rulers. The continuous line sidesteps the question by staying on a single ruler throughout.

The old base squeezed the whole service economy, rents, health, education, transport, phones, restaurants, gold, into a single 28-per-cent "miscellaneous" block. The 2024 rebase splits that block open along the international COICOP-2018 lines, and for the first time each service division gets its own reading. Read the strip as a league table of what is actually firming the new basket.

One division sits in a different country from the rest. Personal care, which carries gold, runs at 15.2%: gold is set in world markets and moves to none of the logic that governs the other eleven. Treat it as a spectator; the basket's story lives in the cluster below.

The rural-urban wedge on the retired 2012 base (§10) was mostly the food wedge, amplified by eighteen points of extra food weight in the village basket. The rebase redraws both baskets with 2024 weights: and the wedge survives it. The village line runs above the town line in seven of the new base's eight prints.

The headline plate above says the village runs hotter than the town. This one says where. Each row is one of the twelve divisions, priced separately in rural and urban markets with each sector's own weights. A row where the two dots sit far apart is a part of the household budget where the village and the town are living through different years.

One rung down from the twelve divisions: the forty-three groups of the international COICOP classification. Cereals, milk, electricity, garments, motor fuels, gold: each a panel on its own short clock. Eight months cannot show a cycle, so the lines are not the point; the ordering and the colour are. Read the grid as a ranked field: panels at the top run hot, panels at the bottom run cool, and the blue marks every group above the all-groups headline.

An index this size is an argument between 358 prices, and the headline is the truce. These are the items themselves, rice to train fares to gold, ranked by the latest print. The top of the table is where a household feels inflation before the statistics do; the bottom is where prices are falling while the average firms. Thirty rows are shown; the field ships complete in the CSV.

Each row carries two different clocks. The y/y column counts the last twelve months. The level column counts from the 2024 average, which the base year fixed at 100: and because that average is barely behind us, the level is the item's whole life on this ruler so far. An item at 130 has risen three times as much as one at 110, whatever this month's y/y says.

Each line answers one question: how much dearer is this part of the basket than twelve months ago? The blue line is their weighted average, so it must always live between its parts: food supplies the swings, core supplies the level, fuel supplies occasional noise.

The shaded band is law, not decoration. Since 2016 the RBI must explain itself to government whenever the headline spends three straight quarters outside 2–6%. Every excursion above the ceiling, 2019–20, 2022, mid-2023, triggered that letter. The 2025 prints pose the opposite question: three straight months below the floor.

The most violent line in the guide, and nearly half the basket. Spikes are usually weather, not policy: failed monsoons surface here three to six months late, and gluts push the line below zero, as in 2018–19 and again in late 2025. Negative food inflation is relief at the shop and distress on the farm, in the same figure.

Only loosely tied to crude: excise duty, LPG subsidy accounting and electricity tariffs dominate. A flat stretch means stable policy, not stable world prices. It moves the headline loudly and means little by itself: never infer the global energy market from this line.

Core strips out weather and energy, the parts monetary policy cannot reach, leaving rents, health, education, transport services, gold: the demand side of the economy. It moves in long arcs, and its Indian floor is structural; the series has never sustainably visited 2%.

Food and fuel get the headlines; these four groups were half the old basket by weight. Miscellaneous alone, at 28 per cent, held every service the household buys: doctors, school fees, bus fares, phone bills, haircuts, gold. Housing is rent, and the NSO repriced it only every six months, which is why its line moves in steps. These are the slow groups. They rarely make news, and they set the floor under the headline.

The six groups are averages, and averages keep secrets. These are the dials beneath them: every sub-index the NSO publishes, one rule per panel: a grey line, ten years, the same shaded band. The band is the tell. Where it looks wide, the panel lives a calm life; where it thins to a sliver, the panel needed room for swings of thirty points and more.

Vegetables and pulses inhabit a different mountain range from education and health. A single shared axis would be honest about the first and silent about the second, so the grid uses two, and prints both. Compare shapes within a half; compare levels only through the numbers.

Year-on-year inflation is a short memory. This ledger is the long one: each item's index in the old base's last month, against the 2012 average fixed at 100. An item at 250 cost two and a half times its 2012 price; one at 90 got cheaper. The spread is the decade's real story of who paid more for what.

The combined index is not a third survey; it is these two weighted together. But the baskets are built from different lives. Food weighs 54.18 in the rural basket against 36.29 in the urban one, and housing, 21.67 of the urban index, does not exist in the rural index at all, because there is no rent market to survey in most villages.

The consequence is mechanical. When food is dear, the village line rides above the town line; when food is cheap, it dives below. The wedge between the two is mostly the food wedge, amplified by eighteen points of extra weight.

The rebase reached the states with the national index: every state's General index restarts at a 2024 average of 100, and each month from January 2026 adds one same-base reading. This strip reads exactly like the retired one below it, the ends carry the news, the middle is a crowd, but the two cannot be compared number for number. A state printing low on the old ruler in December and high on the new one in June did not surge; the measure changed, and six months passed, in unknown proportion.

The count also changed. The old scheme published thirty-five state indices; the new base serves thirty-six: Arunachal Pradesh and Ladakh appear as their own series for the first time, and the merged Dadra & Nagar Haveli and Daman & Diu, two rows on the old ruler, is now one.

The national number is an average of thirty-six. This table opens it the other way: one row per state, one column per part of the household budget. Read along a row to see what drives prices where you live; read down a column to see whether, say, health costs are rising everywhere or only in a few places. A deeper tint means faster price rises; grey marks prices that fell over the year.

Each state has its own basket, so a cell compares a state with its own past, not with another state. The rural and urban versions of this table, and the forty-three groups by state, are in the data downloads and the every-cut table.

The tolerance band is an all-India law; no state is bound by it. Each state index prices its own basket, more fish in Kerala, more wheat in Punjab, so what is shown is the published General index of each state, not a recomputation. Dispersion is the normal condition. A single national rate is always an average of genuinely different price experiences.

Read the ends, not the middle. The middle of this strip is crowded and interchangeable; the outliers are where something specific, a crop, a tariff, a supply line, is happening.

Both lines are rural, but they describe different people. The grey line is the rural half of the national CPI: every rural household, weighted by how it spends. The blue line is the agricultural labourer alone: a poorer household, whose basket the Labour Bureau weights far more heavily toward food, and whose index reprices the floor wages that MGNREGA and state rate-cards pay.

Because food dominates the labourer's basket, the blue line is the more violent one: it rises faster in a bad monsoon and falls further in a glut. Over the decade they part by 1.4 pt on average, and by as much as 4.2 pt, Aug 2019 in a single month.

The national CPI describes an average household. This one describes a specific and poorer one: the agricultural labourer, whose index the Labour Bureau has compiled to reprice rural minimum wages: MGNREGA rates and state wage floors move with it. Its basket is far heavier in food than the national CPI's, so it answers to the harvest before anything else.

The base year stood at 1986-87 for nearly four decades, the longest-serving ruler in Indian statistics, until the 2019=100 series took over from June 2025. The old and new series price different baskets, from different surveys of a different countryside. They are shown side by side, and the seam is drawn, not smoothed.

Two indices, one countryside. The rural labourers' basket contains the agricultural labourers' almost entirely, so the lines move together and part by fractions of a point. The gap, small as it is, is still information: AL above RL means the pressure sits in the farm household's own basket rather than the village's.

Food dominates this basket, so the food line is close to being the index; the other four groups are the supporting cast. Watch fuel & light separately: firewood and kerosene, not LPG cylinders, set its rhythm here, and it answers to a different market than the urban fuel bill.

One countryside, many inflations. Each state's agricultural-labour index prices its own harvest, more rice in the south and east, more wheat in the north, so the spread is genuine, not noise. There is no RBI band here: the tolerance law binds the all-India CPI, not the farm labourer's basket.

The new base covers thirty-four states and union territories, against twenty on the old one, and prices two households in each: the agricultural labourer and the wider rural labourer, who might be a mason, a weaver or a cart-puller. The two dots in a row are the same village markets read with two baskets. Where they sit apart, the farm worker's heavier food share is pulling one way.

This is the oldest continuous price index India keeps for a group of people, and the one with the most money riding on it. The dearness allowance of every central government employee and pensioner is set from it twice a year, and so are wages in many factories, mines and plantations. A point on this index is rupees in a few crore pay packets.

The household it describes is a worker in one of the Bureau's industrial centres: a mill town, a mining belt, a port, a tea garden. The basket was redrawn four times, in 1960, 1982, 2001 and 2016, each from a fresh survey of what such families buy. Each colour is one basket. Where one ends and the next begins there is a fine rule, not a join: the Bureau publishes a linking factor to carry wages across, but a growth rate from one basket to the next would compare two different shopping lists.

Six groups, as in the retired national CPI, and the same lesson: food moves the index. The group to watch here is housing. The Bureau prices rents in the industrial centres twice a year, so the housing line climbs in steps rather than drifting, and a step in January or July can move the whole index in a month when food is quiet.

Each dot is a town where the Bureau sends price collectors to the markets an industrial worker uses. The centres are chosen for their factories and mines, not to cover a state, so read them as places: Jamshedpur, Coimbatore, Kolkata. Dearness allowance is paid on the All-India figure, so a worker in a centre well above the All India rule is falling behind on an allowance that averages the country.

A price index only tells half the story for a wage earner. The other half is the wage. The Labour Bureau collects both for rural India: what a day's work pays in each occupation, and what the farm labourer's household pays for its basket. When the blue line runs above the gold one, a day's wage bought a little more than it did a year before; below it, a little less.

The wage is the average man's daily wage across the Bureau's sample villages, for the general agricultural labourer: watering, weeding, the work anyone can be hired for. It is not what any one worker earns, and it says nothing about how many days of work there were.

The same day's work, priced in twenty states. These are rupees, not an index, so states can be compared directly, with one caution: a rupee buys more rice in some states than others, and this chart does not adjust for that. The other 24 occupations, from ploughing to plumbing, are in the data downloads and the every-cut table below.