How Oil Prices Reach India's Inflation, Rupee and Markets
An oil move reaches an oil importing country by two routes at once. The move raises the import bill, and that bill is settled in somebody else's currency. A larger bill is itself a demand for that currency. So the price move and the currency move arrive together, they multiply rather than add, and what finally lands in a price index is a small fraction of where the chain started.
Why a commodity price swings the way it does, and why fresh supply takes so long to answer a shortage, are settled where the commodity cycle is covered. The definition of a supply shock, and what separates one from a change in demand, is settled where shocks are covered. The exchange rate itself, how it is quoted and what pushes it about, is settled where the external account is covered, along with the record that trade flows are written into. How a basket of prices becomes one headline reading, and how much of a cost increase turns into a price increase, are settled under inflation and prices. Every one of those is borrowed here. The chain itself runs end to end, from a barrel to a price index, with the arithmetic worked step by step and every intermediate figure printed rather than summarised.
All the numbers below belong to the Republic of Sankhya and its partner economy Marut, a pair of economies made up for teaching. Sankhya buys oil priced in Marut units and pays for it in rupees.
What are the two routes out of one oil move?
Start with something concrete. Consider a buyer saving for an imported camera priced abroad. Two separate things can make it cost more this month than last month. The camera itself got dearer where it is made, or the buyer's money now buys less of the money it is priced in. Both show up as one bigger number on the bill, and neither one is the other. Looking only at the final rupee figure, the buyer cannot tell which of the two happened, and certainly cannot tell whether both did.
An oil importing country is that same buyer at national scale, and the two routes have names. The first route is the bill: barrels bought, times the price of a barrel, and nothing else in it. It is pure arithmetic on a quantity and a price, and with both numbers in hand the answer is exact. The second route is the currency. The seller does not want rupees, so Sankhya cannot pay for oil in rupees. Sankhya must first buy Marut units with rupees and then hand those over. So the oil bill is not only a payment, it is an instruction to go into the currency market and buy a certain quantity of somebody else's money.
The second route is created by the first, so the two are not independent events that happen to coincide: a larger bill in Marut units is, by construction, a larger purchase of Marut units. That is why they arrive together rather than one after the other, and it is why treating them as two separate stories with two separate timings gets the shape of the whole thing wrong.
One honest qualification before the arithmetic starts. A currency moves for many reasons, and the oil bill is only one of them. Interest rate differences, capital flows, and what the rest of the world is doing all push on the same rate, and the flow that goes into the external account is written up where the external account is covered. The exchange rate is therefore not derived from the oil bill anywhere below. The currency move is taken as given, exactly as a reader would find it, and what follows shows what happens when both moves land on the same barrel.
An oil move reaches a domestic price by two routes. Which pair is it?
Why are the two routes not independent of each other?
Why do the two moves multiply instead of adding?
Four numbers carry this part, and they are the only four it needs. A barrel of Sankhya's oil starts at 50.00 Marut units. One Marut unit starts at 80.00 rupees. So a barrel starts at Rs 4,000/-. Now the oil price rises to 60.00 Marut units, a move of 20.00 per cent, and the Marut unit rises to 84.00 rupees, a move of 5.00 per cent. A barrel now costs 60.00 times 84.00, or Rs 5,040/-. Against Rs 4,000/-, that is a rise of 26.00 per cent.
Almost nobody guesses 26.00. The instinct is to add 20.00 and 5.00 and call it 25.00, and being one point out on a single barrel feels like nothing worth arguing about. Taken slowly and in order, the two steps show where the extra point comes from. Move the oil price first: 60.00 Marut units at the old rate of 80.00 is Rs 4,800/-, or Rs 800/- more than Rs 4,000/-, a clean 20.00 per cent. Now move the currency. The currency move adds 5.00 per cent, but 5.00 per cent of what? Of Rs 4,800/-, not of Rs 4,000/-. The answer is Rs 240/-, and Rs 240/- is Rs 40/- more than the Rs 200/- the currency move would have produced had the oil price never moved.
The extra Rs 40/- is the currency move applied to the price rise, and it is the whole of the missing percentage point: it never cancels, it does not depend on which step comes first, and it grows as either move grows. Run in the other order, the two steps land in exactly the same place. Move the currency first and a barrel is 50.00 times 84.00, or Rs 4,200/-. Then move the oil price: 20.00 per cent of Rs 4,200/- is Rs 840/-, and Rs 4,200/- plus Rs 840/- is Rs 5,040/-. Same answer, and the same Rs 40/- sitting inside it.
The gap has a shape worth remembering: it is the two moves multiplied together and divided by a hundred. Twenty times five over a hundred is one, the point just found. Push both moves up, though, and the gap stops being a rounding curiosity. If oil moved 50.00 per cent and the currency 15.00 per cent, adding gives 65.00 per cent. Multiplying gives 1.50 times 1.15, or 72.50 per cent. The gap is 7.50 percentage points, and fifty times fifteen over a hundred is exactly 7.50. The error from adding is small when both moves are small and it gets steadily worse in the situations where somebody is actually paying attention.
One point on a barrel is also not nothing once it is multiplied by the barrels. Rs 40/- per barrel across the 36.5 crore barrels Sankhya buys in a year is Rs 1,460 crore. Rs 1,460 crore is the size of the mistake, in money, from choosing the wrong arithmetic operation.
Oil rises 20.00 per cent in Marut units and the Marut unit rises 5.00 per cent against the rupee. What is the move in the rupee cost of a barrel?
Why is the answer not 25.00 per cent?
Where does the bigger bill actually land?
Now put the barrel back into a country. Sankhya burns 36.5 crore barrels a year, or exactly 0.1 crore barrels a day. Before the move, at Rs 4,000/- a barrel, the annual oil bill is Rs 1,46,000 crore. After the move, at Rs 5,040/- a barrel, the same 36.5 crore barrels cost Rs 1,83,960 crore. The increase is Rs 37,960 crore.
Two comparisons make that number mean something. Measured against Sankhya's nominal outputThe total value of everything an economy produces in a year, counted at the prices actually charged during that year rather than adjusted for price changes. Where the figure comes from and how it is assembled is settled where national income is covered. of Rs 17,47,200 crore, the increase is 2.17 per cent. Measured against goods imports of Rs 3,85,000 crore, the oil bill before the move was already 37.92 per cent of everything Sankhya bought from abroad in goods, leaving Rs 2,39,000 crore, or 62.08 per cent, for everything else.
Because oil is 37.92 per cent of the whole goods import bill, a move of this size does not sit beside the external account, it sits inside it. An extra Rs 37,960 crore has to be paid for out of what Sankhya earns abroad, out of what it borrows abroad, or out of reserves, and which of those it comes from is a question about the current accountThe running record of what a country earns from and pays to the rest of the world on trade, services and income, as distinct from the borrowing and investment flows that finance any gap. The record it keeps, and how a gap is financed, is settled where the external account is covered. rather than about oil. Sorting between those three belongs where the external account is covered. The arithmetic here does establish one thing: a reader who treats the oil bill as a separate item to be considered after the trade balance has been discussed has misplaced more than a third of the trade balance.
One assumption is doing quiet work here and it deserves to be said out loud. Holding the barrels constant at 36.5 crore is the short run assumption. Over a longer stretch, quantity responds: some fuel gets used more carefully and some gets replaced. The response is slow, and that slowness is precisely the point made where the commodity cycle is covered. Holding quantity fixed is the honest way to isolate what a price move does on its own.
The oil bill rises by Rs 37,960 crore. Sankhya's output is Rs 17,47,200 crore. What share of output is that increase?
How much of a 26.00 per cent move reaches a price index?
Most readings of an oil headline fall apart here, and they fall apart in a very specific way. Two steps stand between a higher input cost and a higher inflation reading, and both of them shrink the number.
The first step is the basket weightThe share of a typical household's total spending that one item accounts for inside a price index. How a basket is assembled, surveyed and weighted is settled under inflation and prices.. Fuel and the things closest to it are 8.00 per cent of Sankhya's basket, so 92.00 per cent of what a household spends is on something else entirely. Even if every single rupee of the oil move reached the shelf, the index would only feel 8.00 per cent of it: 26.00 per cent times 8.00 per cent is 2.08 points.
The second step is pass-through. Somebody in the chain absorbs part of a higher cost, or delays it, or was already carrying stock bought at the old cost, so not every rupee reaches a shelf price. At a pass-through share of 60.00 per cent, the 2.08 points becomes 2.08 times 60.00 per cent, or 1.25 points. Every step of the chain loses part of the move, so an account that carries 26.00 straight through to an inflation figure has skipped two steps and overstated the answer by a factor of 20.80.
Notice the units, and notice them before going further. The 26.00 is a per cent: it describes how much one price changed. The 2.08 and the 1.25 are points: they describe how much one reading of a whole index is lifted. A number reported without its unit here is not a small imprecision, it is a completely different quantity, and the same trap runs through this whole subject in different clothes: reserves get quoted in days of use in one place and months of imports in another. Say the unit every time, even when it feels laboured.
A 26.00 per cent domestic move, at an 8.00 per cent basket weight and 60.00 per cent pass-through, contributes what to a price index?
The whole chain in one table
Every step sits in one place below, with its own unit on every line. Nothing in the table is asserted: each row is either an input or the row above it multiplied by something visible. A chain a reader cannot recompute is a chain they have to take on trust, and taking a chain on trust is exactly how a wrong number survives.
| Step | What it is | Before | After | The move |
|---|---|---|---|---|
| 1 | The price of a barrel, in Marut units | 50.00 | 60.00 | up 20.00 per cent |
| 2 | Rupees to one Marut unit | 80.00 | 84.00 | up 5.00 per cent |
| 3 | The rupee cost of one barrel, step 1 times step 2 | Rs 4,000/- | Rs 5,040/- | up 26.00 per cent |
| 3a | What adding the two moves would have given instead | Rs 4,000/- | Rs 5,000/- | up 25.00 per cent |
| 4 | Barrels bought in a year, held constant | 36.5 crore | 36.5 crore | no change |
| 5 | The annual oil bill, step 3 times step 4 | Rs 1,46,000 crore | Rs 1,83,960 crore | up Rs 37,960 crore |
| 6 | The increase against output of Rs 17,47,200 crore | not applicable | not applicable | 2.17 per cent |
| 7 | The oil bill against goods imports of Rs 3,85,000 crore | 37.92 per cent | not applicable | not applicable |
| 8 | Step 3 at the 8.00 per cent basket weight, full pass-through | not applicable | not applicable | 2.08 points |
| 9 | Step 8 at 60.00 per cent pass-through, what arrives | not applicable | not applicable | 1.25 points |
Read the last column downwards and the shape of the whole thing is visible in one glance. The column starts at 20.00 per cent, grows once to 26.00 per cent because two moves multiplied, and then shrinks twice, first to 2.08 points and then to 1.25 points. The chain grows once and shrinks twice, and a reader who remembers only that will already be more careful than most.
Run the whole chain yourself, with the wrong answer drawn beside the right one
The controls set the oil price move, and then the currency move, the basket weight and the pass-through share. The panel runs every step and prints every intermediate figure, so the arithmetic can be checked rather than accepted. In the top block the answer from adding the two moves is drawn as its own bar right above the answer from multiplying them, and the shaded strip between the two is the gap. The gap stretches as either move is pushed further. The panel opens on the case traced above: 26.00 per cent, Rs 1,83,960 crore and 1.25 points.
What reaches a business, and where the chain stops
A business that burns fuel sees a higher cost. Whether that higher cost reaches its profit, and by how much, depends on two things that vary enormously from one business to the next. The first is cost shareThe proportion of a business's total costs that one input accounts for. A haulage operator and a software office can face the same fuel price and be in entirely different positions because the share differs. How costs are read off a set of accounts is settled where financial statement analysis is covered.: what proportion of everything the business spends goes on fuel. The second is pricing powerHow much of a cost increase a business is able to hand to its own customers without losing them. Pricing power depends on what else the customer can buy instead, and on whether any price is fixed by contract. Where it comes from is settled where competitive position is covered.: how much of the increase it can hand to its own customers.
Think of two businesses on the same street. A haulage operator running twelve lorries spends a large slice of everything it earns on diesel, and its rates for the next eight months are already fixed in a contract signed last year. A tailoring workshop next door pays for fuel only when it sends a delivery scooter out. Same oil move, same street, same day, and the two are simply not in comparable positions. Neither is the operator's situation a verdict: it depends entirely on what the contract says and how long it runs.
Naming the channel is as far as the arithmetic reaches. No margin, share price, index level or direction of travel follows from a cost share and a pricing power alone, for Sankhya or for anywhere real. The reason for stopping is not squeamishness. Somebody sells the oil, so the same move that raises one business's costs raises another business's revenue. Adding up who wins and who loses across a whole economy is a different exercise with a different method, and doing it badly by generalising from one channel is worse than not doing it at all.
Which parts of this chain are arithmetic and which are judgement?
The most useful step in the whole chain takes two minutes to learn. The chain sorts into two boxes, and the sorting comes before any argument about it.
Into the first box goes everything that is settled once the inputs are given. The two moves multiply to 26.00 per cent, not 25.00. A barrel goes from Rs 4,000/- to Rs 5,040/-. The annual bill reaches Rs 1,83,960 crore, an increase of Rs 37,960 crore. The increase is 2.17 per cent of output of Rs 17,47,200 crore. And 26.00 per cent at an 8.00 per cent weight is 2.08 points. Nobody can disagree with any of that without disagreeing with multiplication. Sitting just alongside the arithmetic is one thing that is measured rather than computed: the 8.00 per cent weight itself. Somebody counted that weight through a spending survey and published a method for it. The method can be questioned, and it can also be read.
Into the second box goes everything nobody can settle in advance. How much of the higher cost actually reaches a shelf price. How long the move lasts, and whether it reverses, the question of persistenceWhether a price move stays in place or unwinds, and for how long. A move that reverses within a quarter and a move that stays for two years have very different consequences for anything that is measured year on year. How a price index treats a level that stays put is settled under inflation and prices.. Whether this is a shock or a turn in the cycle. How much any given business can hand to its own customers. And what the currency would have done anyway, for reasons that have nothing to do with oil.
Every real disagreement about oil and inflation lives in that second box, and separating the two boxes is what lets two people argue about the right thing instead of throwing whole final numbers past each other. The practical payoff is immediate. When somebody claims an oil move adds a certain amount to inflation, exactly two questions follow, and neither of them is rude: which pass-through share is being used, and how long the move is expected to last. If the first has never been thought about, the disagreement has been found in under a minute.
Which of these is a judgement rather than arithmetic?
The error that gets made, and why the answer still looks sensible
A reader sees a headline saying oil is up 20 per cent. The reader reports the domestic cost of oil as up 20.00 per cent, then remembers, correctly, that fuel is only 8.00 per cent of the basket, applies that, and reports a contribution of 1.60 points to inflation. The working sounds careful. In fact it is two mistakes stacked on top of each other.
The first mistake understates. The currency moved as well and the two multiplied, so the domestic move was 26.00 per cent, not 20.00 per cent. Using 20.00 leaves 6.00 percentage points on the floor. The second mistake overstates. The pass-through step was never applied at all, so the whole cost increase was treated as reaching the shelf when only 60.00 per cent of it does.
Here are all four answers, each one being the domestic move times the 8.00 per cent basket weight times whatever pass-through share was applied. Both mistakes together gives 1.60 points. Only the pass-through mistake fixed gives 0.96 points. Only the move mistake fixed gives 2.08 points. Both fixed gives 1.25 points, the answer.
The two mistakes run in opposite directions and only partly offset each other, so 1.60 points looks like a careful answer while being reached by two errors, and a final number that looks reasonable says nothing at all about whether the working behind it was right.
The fix is two habits and no more than that. Multiply the moves, never add them. Then attenuate in order: weight first, pass-through second. And publish the intermediate figures, the only reason the mistake above is visible at all.
A reader takes a 20.00 per cent oil headline, reports the domestic rise as 20.00 per cent, applies only the 8.00 per cent basket weight and reports 1.60 points. What went wrong?
What does somebody actually do with this at a desk?
Take the household first. Everybody has one and the arithmetic is the same arithmetic. No index is needed to run it. The fraction of a month's spending that goes on fuel and transport serves as a household basket weight. On spending of Rs 30,000/- a month with fuel at 6.00 per cent of it, that is Rs 1,800/-. A 26.00 per cent move on that line is Rs 468/- a month. Rs 468/- is a real number and worth planning around, and it is also nothing like a 26.00 per cent rise in the cost of the month. The household itself is best placed to compute that weight, and doing it once takes five minutes and permanently changes how a fuel headline reads.
An equity analyst covering a business that burns fuel does the sorting before the sizing. First the sign: does this business pay more or receive more? Getting the sign wrong cannot be recovered later. Getting the size roughly right is only imprecise. Then the sizing, and it needs three things rather than one: what fraction of the cost base the fuel is, how much of any increase the business can hand on, and how much of the move was already reflected in a traded price before the headline appeared.
A treasury team inside an importing business runs the chain forwards rather than backwards. Such a team already knows its barrels, or its tonnes, or its kilowatts. The reminder it needs is that one exposure is really two wearing one coat: the commodity price and the currency, arriving multiplied. Managing one and leaving the other alone leaves the interaction term entirely uncovered, and the interaction term is the part that grows fastest when both move together.
A reader of policy commentary uses the chain as a scale check. If a headline inflation reading is 6.70 per cent and this oil move contributes 1.25 points of it, then oil accounts for 18.66 per cent of that reading and everything else accounts for the remaining 5.45 points. Both components are stated above, so the division can be checked. The split establishes nothing about what happens next, and no arrangement of these numbers will.
What is needed before drawing any conclusion
The honest end of this chain is a shorter list of things still unknown, so what follows is four questions rather than a verdict.
Is this a shock or a turn in the cycle? A supply event that reverses and a cycle that is turning look identical on the first day and behave nothing alike afterwards. Both are settled elsewhere: shocks where shocks are covered, the cycle where the cycle is covered. How long is the move expected to last, and expected by whom? A move that unwinds inside a quarter contributes to one reading and then takes it back out. What did the currency do on its own? If the rupee was moving anyway for reasons that have nothing to do with oil, then part of the 26.00 per cent was going to happen regardless and attributing all of it to oil overstates the oil story. And how much imported contentThe share of a finished price that was bought from abroad rather than made at home. A shirt sewn locally from imported cloth carries imported content even though nobody imported a shirt. An import price can therefore reach a price index through goods that are not themselves imports. sits in the particular basket under examination? An 8.00 per cent fuel weight is not the same as an 8.00 per cent imported content, and a basket full of locally made goods that are moved around by lorry has more oil in it than its fuel line suggests.
Nobody can forecast where an oil price, a currency, a price index or a market goes next. The chain multiplies and attenuates a move that has already happened, and it carries no machinery at all for the next one. A forecast cannot be checked. A chain with every link visible and every intermediate figure printed can be, and that is what makes it worth more.
Where an Indian reader would go to check any of this
Two benchmark crudeA particular grade of oil from a particular place whose quoted price is used as the reference point for pricing other cargoes. A benchmark grade, and how two of them come to differ, is settled where crude benchmarks are covered. grades, Brent and West Texas Intermediate, are the reference points most often quoted in commentary, and India maintains a strategic petroleum reserve held in underground storage.
For the quantities and values of India's crude imports, the Ministry of Petroleum and Natural Gas and its Petroleum Planning and Analysis Cell publish the series. For the consumer price index and the item weights inside its basket, the Ministry of Statistics and Programme Implementation publishes both the readings and the method. For the exchange rate and the external account, the Reserve Bank of India publishes the record. For merchandise trade with petroleum separated from everything else, the Ministry of Commerce and Industry publishes the monthly figures.
A figure written from memory in this subject goes stale quietly and gives no sign that it has.
References
| Body | What it publishes | Site |
|---|---|---|
| Ministry of Petroleum and Natural Gas, Government of India | Through its Petroleum Planning and Analysis Cell, the quantity and value of crude India imports. Named here because that is where a real barrel count would have to come from | mopng.gov.in |
| Ministry of Statistics and Programme Implementation | The consumer price index, and the weight each item carries inside the basket. The 8.00 per cent used above is an assumption for Sankhya and not a reading taken from anywhere | mospi.gov.in |
| Reserve Bank of India | The exchange rate record and the external account into which an import bill is written. Named for where the currency half of this chain is measured | rbi.org.in |
| Ministry of Commerce and Industry, Government of India | Monthly merchandise trade, with petroleum reported apart from the rest of goods. The one place the 37.92 per cent question can actually be answered for a real country | commerce.gov.in |
| International Monetary Fund | Research on how commodity price moves travel into prices, output and external balances. Where a measured pass-through estimate would have to come from, since none is offered here | imf.org |
The Republic of Sankhya, its partner economy Marut and the Marut unit are invented.
Educational material. Not advice on any investment, tax, budget or market position.
