The Components of Indian Inflation: Food, Energy and Core
Inflation is a sustained rise in the general level of prices. One price rising is not the same thing. Inflation comes from spending outrunning what an economy can produce, from costs rising and being passed along, or from money buying less abroad. Each of those causes behaves differently, and a single headline figure contains all three at once, so the first job is to tell them apart.
Nobody can observe the general level of prices directly. Millions of separate prices exist, and the general level is built out of them by choosing which prices to watch and how much each should count. The choosing is why one economy in one year can honestly produce several different inflation numbers at once.
Three routes start a general rise, a headline can be taken apart into the parts that built it, and core is that same headline with named parts left out.
Rebuild a headline and a core reading from a basket entered by hand
The fields take the group weights and the group price changes off an index release, one month of household spending, and a choice of which groups this particular core reading drops. Everything below recomputes as each entry changes. The calculator opens on the Sankhya year worked through in the rest of this guide, so leaving every field alone reproduces that example exactly.
| Group | Weight | Its own prices | Weight times price change | In this core reading |
|---|---|---|---|---|
| Food | ||||
| Energy | ||||
| Everything else | ||||
| Headline, the three contributions added | 100 per cent |
Left alone the instrument reproduces the case this guide works through. Food at 40 per cent of the basket rising 12.00 per cent contributes 4.80 points, energy at 10 per cent rising 9.00 per cent contributes 0.90, and everything else at 50 per cent rising 2.00 per cent contributes 1.00, so the headline is 6.70 per cent against a core of 2.00. A household spending Rs 20,000/- a month pays Rs 21,340/- a year later.
In the instrument above, set the core reading to drop nothing at all and leave every other field alone. The core reading moves to 6.70 per cent and the distance between the two readings goes to nil. Why?
What actually causes inflation, and why is one price rising not enough?
A shopkeeper puts up the price of onions and somebody says inflation is back. But onions became scarce in one district after a failed crop, so anyone who wanted them gave up more of everything else, and the next harvest brings the price down. The onion move is a relative price changeOne price moving compared with other prices, usually because that particular thing became scarcer or more plentiful. The move says something about that market and nothing about money.. It happens every day.
Inflation is a different event: the whole level moving. Rice, bus fares, school fees and rent all cost more, at their own speeds but in the same direction. One price rising and the general level rising have different causes, different durations and different cures. Confusing the two is the root of most bad reasoning in this subject.
Three routes make a general level move: spending running ahead of what can be produced, costs underneath everything rising and being passed along, or the money buying less abroad. Every inflation story is one of those or a mixture. The word sustained in the definition is doing real work: all three routes have to keep going to keep prices going. A one off jump in a price level is arithmetic that lands once and then stops appearing in the comparison; inflation is a jump that keeps being repeated.
Onions triple in price after a failed crop while nothing else in the shop moves. What has happened?
What is demand-pull inflation, and how does spending outrun what can be made?
Demand-pull is the version most people picture. Total spending, what an economist would call aggregate demandThe total of everything households, businesses, government and foreign buyers together want to buy in an economy over a period, added up as one quantity instead of market by market., rises faster than the economy's ability to make things. Since the quantity cannot stretch in the short run, the price does the adjusting.
Feel it at household scale. A wedding season begins, four hundred households want the same eleven caterers in the same three weeks, and no caterer can conjure a twelfth kitchen by Friday. The quotes go up for everybody, including the household that would happily have paid last year's rate.
How a price settles when buyers want more than sellers can supply is not re-derived here. At the level of a whole economy the constraint becomes the country's entire productive capacity, and pressure against it leaks sideways: the caterer who is fully booked bids harder for cooks, and cooks then cost more for the canteen that had nothing to do with any wedding. Demand-pull spreads because the constraint being hit is shared.
What is cost-push inflation, and where does it begin instead?
Turn the picture around. Nobody wants more than last year, but something sitting underneath almost everything gets dearer: diesel, electricity, imported steel. Producers pass what they can of it into the price on the shelf. Cost-push can begin in an input costWhat a producer pays for the things it needs in order to make what it sells: raw material, fuel, power, transport and bought in components, as distinct from what it charges at the end. nobody thinks of as a consumer price.
The street vendor version is exact. A vendor selling idlis pays for rice, dal, gas and the auto that brings the vegetables in. When the cylinder goes up the queue did not lengthen and nobody asked for anything new, but the plate has to carry the cylinder.
Demand-pull vs cost-push inflation: what single observable actually separates them?
In the price index itself the two are indistinguishable. Prices went up. The index does not record why, and no amount of staring at the figure will reveal which engine produced it. The index was never built to answer the question.
The one thing that differs is what output is doing alongside. Under demand-pull, buyers press against capacity, so output rises with prices; under cost-push, producing got dearer, so output is squeezed while prices climb. Demand-pull arrives with output rising and cost-push arrives with output falling, and that opposite movement in the quantity is the one observable that separates two causes which look identical in the price column.
The third difference is durability. Cost-push from one event is a level shift wearing the costume of a rate: the cylinder went up once, the plate is dearer for twelve months, and then the comparison rolls past the increase. A one off cost shock stops showing up in the inflation rate about a year later without a single price ever falling. An inflation rate that is falling is therefore not the same statement as prices that are falling.
Two economies both report prices rising. In one, output is rising as well; in the other, output is falling. What does that show?
What is food inflation, and why does food inflation dominate a developing basket?
Between year 2 and year 3 in the Republic of Sankhya, an invented country, food prices rose 12.00 per cent, energy 9.00 per cent and everything else 2.00 per cent. Food is 40 per cent of what a Sankhya household spends, energy 10 per cent and everything else the remaining 50 per cent.
Food matters more in Sankhya than in a rich country for one unglamorous reason: the basket weightThe share of a household's total spending that goes on one group of items. A group with a weight of 40 per cent contributes forty hundredths of whatever its own prices do to the overall figure.. Spending four rupees in every ten on food gives food four tenths of a vote in the overall figure, and richer households spend a smaller share on eating. The identical harvest failure produces a much larger inflation reading where food takes a big share of spending. Food is a separate line in developing economies for exactly that reason.
The second feature is what food prices do rather than how much they count. Harvests, rainfall, pests and storage are events rather than trends, and events reverse. Movement that reverses is volatilityHow far and how fast a number swings around, in both directions. A volatile series can rise sharply and then fall back to roughly where it started, unlike one that drifts steadily one way.. A vegetable at Rs 20/- a kilo goes to Rs 60/- after a washed out crop and comes back to Rs 22/-, a two year rise of 10.00 per cent. A haircut walked from Rs 100/- to Rs 104/- and never went down.
Two prices can arrive at almost the same place over two years while one travelled through a tripling and a collapse and the other simply walked, and that difference in route rather than destination is why food is pulled out and looked at separately. A reading dominated by a food spike is a statement about weather.
What is energy inflation, and why does energy inflation reach prices that are not fuel?
Energy is the opposite shape to food: a small weight and a long reach. In the Sankhya basket it is 10 per cent of household spending, and a 9.00 per cent rise there contributes only 0.90 points. The 0.90 points understates energy badly.
Energy is not only something a household buys. Energy is something every producer buys before making anything. One tomato passes through a diesel pump, a diesel truck, a cold store and a tempo to the shop. The household that buys it has paid for four purchases of energy without buying a drop of fuel.
Energy enters the price of nearly everything else as a cost before it ever appears as a purchase, so energy has a small basket weight and a large footprint. A fuel increase shows up twice: once in the energy line straight away and again in other lines over the following months. How far it travels is worked through elsewhere. Energy is volatile like food and, unlike food, an input to the whole rest of the basket. Both features together are why energy is separated out.
Food is 40 per cent of the Sankhya basket and food prices rose 12.00 per cent. How many points does food contribute to the overall figure?
What is imported inflation, and why does an exchange rate multiply it?
The third route is the one most explanations leave out. A country buys things it does not make, priced in a currency that is not the rupee. Nobody in Sankhya spent an extra rupee, no Sankhya harvest failed, and the price on the Sankhya shelf still went up.
The landed cost in rupees is the world price times the rupees a unit of foreign money costs, and both can move at once. In year 2, 100 units at Rs 80/- landed at Rs 8,000/-. In year 3 the world price rose 5.00 per cent to 105 units and the currency 6.00 per cent to Rs 84.80/-, landing at Rs 8,904/-.
The landed cost rose 11.30 per cent, the 5.00 and the 6.00 combined by multiplication rather than added. Adding gives 11.00, and the missing 0.30 points is the world price increase paid at the dearer currency too. A country can import a price rise with no domestic condition changing at all, and the exchange rate can add to a world price rise or cancel it entirely.
Sankhya imports an input. The world price is unchanged this year, but a unit of foreign money now costs 6.00 per cent more rupees. What happens to the landed cost in rupees?
What is headline inflation, and how is a headline built rather than measured?
Headline inflation is the whole basket, every group counted at the share of spending it takes. Nobody measured it. The headline was assembled: somebody chose the groups, somebody measured each group's price change, somebody fixed the weights from a spending survey, and somebody multiplied and added.
The assembly is a weighted averageAn average in which some items count more than others. Each item is multiplied by its share before the adding, so a group taking 40 per cent of spending has four times the pull of a group taking 10 per cent.. Multiply each group's price change by its weight and add the results. The product is that group's contribution, measured in points rather than per cent.
| Group in the Sankhya basket | Weight | Its own prices | Contribution |
|---|---|---|---|
| Food | 40 per cent | up 12.00 per cent | 4.80 points |
| Energy | 10 per cent | up 9.00 per cent | 0.90 points |
| Everything else | 50 per cent | up 2.00 per cent | 1.00 point |
| Headline, year 2 to year 3 | 100 per cent | 6.70 per cent |
Read the contribution column before the weight column. Food takes 40 per cent of Sankhya spending but supplied 4.80 of the 6.70 points, or 71.64 per cent of the movement. The headline is that sum and nothing more: three numbers that can be inspected, multiplied by three weights that can be looked up.
What is core inflation, and why is anything taken out of a basket at all?
Core inflation is the same basket with food and energy removed and the remainder read on its own. In Sankhya that leaves the 50 per cent called everything else, whose prices rose 2.00 per cent, so core inflation is 2.00 per cent.
Most people suspect the point is to make the number look better. It is not. Food and energy reverse, so a reading that is high because a crop failed comes back down when the next crop lands. Core answers whether prices are rising in the parts that do not bounce back, and it answers it well precisely because it is missing the parts that bounce.
Now the step almost everyone gets wrong, and it is arithmetic rather than judgement. Core is not the contribution the remaining group made: that was 1.00 point, and core inflation is 2.00 per cent. Half a basket is not a basket, so the 50 per cent is re-weighted to a whole 100 before its price change can be quoted as a rate, and the weight doubles. The re-weighting is why the gap is smaller than the sum of what was removed.
Sankhya reports headline inflation of 6.70 per cent and core inflation of 2.00 per cent. Is core the real underlying rate?
Headline inflation vs core inflation: what is a 4.70 point gap actually saying?
Sankhya reported 6.70 per cent headline and 2.00 per cent core in the same year, from the same basket, compiled by the same people. Neither is a revision of the other and neither is wrong, and they differ by 4.70 points because they answer two different questions.
Headline answers what a household is actually paying, and a household buys food and fuel. Core answers whether the level is moving in the parts that do not reverse. Core is not more real than headline and headline is not more honest than core: one is about what was paid and the other about what is likely to persist.
The size of the gap is a third fact neither measure carries alone. A 4.70 point gap says the movement sits almost entirely in the volatile part of the basket; had the readings been 6.70 and 6.20, the same headline would have meant prices rising broadly. The gap says how broad the price movement is rather than how large.
Two ways to reach 4.70, and why their agreeing confirms nothing
There are two tidy routes to the gap, and it is tempting to treat the match as a check. The first is subtraction: 6.70 less 2.00 is 4.70. The second weights how far the volatile groups ran ahead: 0.40 times the 10.00 points by which food beat core, plus 0.10 times the 7.00 by which energy did. The two products are 4.00 and 0.70.
The two routes are the same equation rearranged, not two independent measurements, so their agreeing proves the algebra and says nothing about whether the weights or the price changes are right. A genuine check comes from outside the arithmetic: a different spending survey, or an index built on a different principle.
An economy reports headline inflation of 6.70 per cent and core inflation of 2.00 per cent. What does the size of that gap show?
Move the three group price changes and rebuild the headline yourself.
The panel opens on the Sankhya year exactly: food up 12.00 per cent, energy up 9.00, everything else up 2.00. The three build a headline of 6.70 per cent against a core of 2.00 and a gap of 4.70 points. Three drawings redraw together whenever a slider moves. The top block shows each group's contribution as a bar running right from zero when prices rose and left when they fell. The middle scale places the headline marker and the core marker on one line, so the gap between them has a visible length. The strip at the foot converts the two rates into what Rs 1,00,000/- of savings would still buy after ten years if that rate ran the whole time. The one thing the year 3 reading cannot show is what happens when food and energy are dragged below zero, until the headline marker slides to the left of the core marker. Headline above core is not the natural order of things, only what happened in this one year.
How inflation erodes purchasing power over time, and why does the damage accelerate?
Purchasing power is what a fixed amount of money will buy, and inflation is the rate at which that shrinks. If prices rise 6.70 per cent, Rs 100/- at the end of the year buys what Rs 93.72/- would have bought at the start. Now do it again next year, on the reduced amount.
The repetition is the point, and it makes the erosion compoundingApplying a rate again and again to a result that already includes every earlier application, so each step works on a different base than the one before rather than on the original amount. rather than adding up. Each year's rise applies to a level that already contains the previous ones.
| What Rs 100/- still buys after | At the headline rate of 6.70 per cent | At the core rate of 2.00 per cent | The gap |
|---|---|---|---|
| 5 years | Rs 72.31/- | Rs 90.57/- | Rs 18.26/- |
| 10 years | Rs 52.28/- | Rs 82.03/- | Rs 29.75/- |
| 20 years | Rs 27.33/- | Rs 67.30/- | Rs 39.97/- |
Read the last column first. The two rates are Rs 18.26/- apart after five years, Rs 29.75/- after ten and Rs 39.97/- after twenty. The erosion compounds, so a gap of 4.70 points opens into a widening distance rather than a constant one.
Put a person in it. The abstraction hides who pays. A retired schoolteacher receives a fixed pension of Rs 20,000/- a month that nobody adjusts, and after ten years at the headline rate it buys what Rs 10,456/- buys today. Inflation transfers value quietly from anyone holding a fixed rupee claim to anyone owing one.
If inflation ran at 6.70 per cent for ten years, roughly what would Rs 100/- still buy at the end of them?
How inflation drives interest rate decisions: what route does a published reading travel?
How a central bank weighs what it sees is treated under monetary policy.
An interest rate is quoted in money, but the usefulness of those rupees depends on what prices did while they were borrowed or lent. A rate quoted in money and a price level in motion are two halves of one question. The two halves together are the entire reason a price statistic ends up in front of the people who set rates.
The drawing below walks the five steps. The drawing cannot say how thin the fourth step is: the distance between the reading and the aim is one input among several, and output, expectations, the currency and the banking system all sit on the same table.
Which reading enters that comparison, a headline or a measure with the volatile groups removed, is a choice written into the arrangement rather than a fact of nature, and the two can point in different directions in the same year. Sankhya's year 3 makes that vivid: 6.70 per cent and 2.00 per cent are both true of the same twelve months.
What is the India inflation targeting framework, and what kind of arrangement is it?
The stated rate and the date it runs from are rewritten whenever the arrangement is renewed, so both belong to the current document rather than to memory.
India's central bankThe public institution that issues a country's currency and sits at the centre of its banking system. India's is the Reserve Bank of India. is the Reserve Bank of India. An inflation targeting framework is a kind of arrangement a central bank works under, and its defining feature is not the number in it but that a number is stated at all, in advance.
Three properties describe the kind of thing it is, and none requires knowing the number. The arrangement is written down and public, in a document with a date on it, and it names a measure and not just a level. And a public aim set against a published reading makes the aim checkable by anybody.
A central bank does not set prices: it acts on conditions that influence them, with a lag. The arrangement is therefore not a promise that the stated rate will be delivered in any period. The arrangement states an aim and usually an obligation to explain a sustained miss.
Who publishes the price numbers in India, and what to take from each of them
Three names are worth carrying. Price statistics in India are compiled under the Ministry of Statistics and Programme Implementation, the Reserve Bank of India reproduces the series and publishes the arrangement it works under, and the Economic Survey retells a year of prices as narrative. Each of the three dates what it issues, so a weight, a rate or a target is read off the issuing document rather than recalled.
What does an inflation targeting arrangement commit a central bank to?
How inflation affects stocks, bonds and gold differently: what is the mechanism?
Three kinds of holding meet a rising price level by three different routes. A fixed money claim, a claim on a business's output and a metal that produces nothing each show what inflation is from a different side.
Start with a fixed claim in money. A bond of Rs 10,000/- carrying a nominal couponA payment fixed in money terms, such as a set number of rupees each year, with the amount written into the contract and not adjusted for anything that happens to prices afterwards. of 7.00 per cent pays Rs 700/- a year whatever the price level does, so prices rising reduces not the payment but what it buys. After ten years at the headline rate that Rs 700/- buys what about Rs 366/- buys today.
Now take a share, which is a claim on a business's real output. The prices it charges are among the prices rising, so the rupees it takes in can rise with the level. The takings can move with the price level only where the business is actually able to raise its own prices, and whether it can is a question about how its customers respond rather than a property of shares in general.
Third, a metal. Gold produces nothing and owes nothing, so there is no coupon to erode and no output whose selling price can rise. Its price is whatever somebody will pay, so the route to a price level runs entirely through what people decide. The three differ in kind and not in degree: one holds a promise fixed in money, one a claim on activity whose prices may or may not move, and one a thing with no cash flow at all.
Interest rates, output, currencies and expectations all move too, and each touches all three cases at once. A mechanism is the route by which one thing reaches another rather than the outcome, and three routes are named here.
A bond pays a fixed Rs 700/- a year and the general price level rises. What happens to that coupon?
What does a lender actually do with an inflation reading?
Not forecast it. Most people assume the opposite. A lending officer takes the reading apart, then asks which of this borrower's prices and costs move with which part of it, and which are nailed down by a contract.
A Sankhya bus operator carries a school run at a fixed fare under a three year contract, so revenue of Rs 24,00,000/- cannot be repriced while its costs move. Against it sit costs of Rs 20,00,000/-, a surplus of Rs 4,00,000/- and repayments of Rs 3,00,000/-, a cover of 1.33 times.
The table applies each group's own Sankhya rate line by line rather than the headline. Costs reach Rs 21,03,000/-, revenue stays where the contract put it, and the cover falls to 0.99 times.
| The bus operator | Before | Its own rate | After |
|---|---|---|---|
| Revenue, fixed by a three year contract | Rs 24,00,000/- | 0.00 per cent | Rs 24,00,000/- |
| Diesel | Rs 9,00,000/- | 9.00 per cent | Rs 9,81,000/- |
| Driver wages | Rs 7,00,000/- | 2.00 per cent | Rs 7,14,000/- |
| Everything else | Rs 4,00,000/- | 2.00 per cent | Rs 4,08,000/- |
| Operating surplus | Rs 4,00,000/- | Rs 2,97,000/- | |
| Cover on Rs 3,00,000/- of repayments | 1.33 times | 0.99 times |
This basket is 45 per cent diesel and holds no food, so the borrower's own cost inflation was 5.15 per cent rather than the headline 6.70. Every borrower has its own basket, and it almost never resembles the household basket that produced the published figure. The damage came from the contract rather than the inflation.
The reader who sees core at 2.00 and decides the headline was exaggerating
Careful people make this mistake. A reader notices 6.70 per cent headline against 2.00 per cent core, remembers that core strips out volatile items, and concludes that the underlying rate is 2.00 per cent and the headline was exaggerating. Every step of that is defensible except the last.
Core is not the rate with the noise removed. Core is the rate with food and energy removed, and in Sankhya food and energy are 50 per cent of what a household spends. A household spending Rs 20,000/- a month puts Rs 8,000/- on food, Rs 2,000/- on energy and Rs 10,000/- on everything else. A year later the same purchases cost Rs 21,340/-, or 6.70 per cent more, exactly the headline. At the core rate the month would have cost Rs 20,400/-, so the readings are Rs 940/- apart in a month and Rs 11,280/- across the year.
A household planning on 2.00 per cent budgets about eleven thousand rupees short, and a lender pricing a loan off core assumes a cost pressure that ignores fuel and food. The calculator above shows both edges of it: with core set to drop nothing the shortfall goes to nil, and with food and energy falling back the household has set aside more than the month cost.
The fix is not to prefer one measure but to name the question being answered. If it is whether the level is moving in the parts that do not bounce back, core answers it; if it is what somebody actually paid, headline answers it. Neither number is the true one, and the habit worth building is naming the question before choosing the measure.
Where does a reader go for a real basket, real weights and a real reading?
Sankhya has no existence, so Sankhya has no statistical agency to check. The issuing document holds what is worth checking: the items an actual basket contains, the weight each carries, and the rate an actual index printed. A weight typed from memory is wrong quietly and stays wrong for years, so every magnitude comes from the issuing document itself.
| Issuer | What to look for there | Site |
|---|---|---|
| Ministry of Statistics and Programme Implementation, and the National Statistical Office inside it | The consumer price index and the wholesale price index: which items sit in each basket, what weight each group carries, how a group is defined, and how a released reading is labelled | mospi.gov.in |
| Reserve Bank of India | Its compiled statistical volumes, and its own published documents describing the arrangement it works under and what that arrangement requires of it, each carrying the date it was issued | rbi.org.in |
| Ministry of Finance | The Economic Survey, which retells a year of prices as narrative and names a source underneath every table it prints, so a figure can be traced back to whoever compiled it | indiabudget.gov.in |
The Republic of Sankhya is invented.
Educational material. Not advice on any investment, tax, budget or market position.
