Supply Shocks: When Output Falls Without Demand Changing

A supply shock is a fall in how much of something can be produced or delivered, with demand left untouched. Less is available and no less is wanted, so the price rises while the quantity used falls. A fall in demand moves the quantity the same way and the price the opposite way, and that one difference is what makes a supply shock the harder of the two to answer.

A supply shock rests on one small mechanism already established: two lines on a chart, one for how much sellers will part with at each price, one for how much buyers will take at each price, meeting at a single price and a single quantity. A supply shock moves one line and leaves the other exactly where it stood, and every awkward thing that follows comes out of that one restriction: the price and the quantity parting company, the lever that cannot win, the shortage that refuses to clear for years.

What exactly moved, and what stayed where it was?

Start with the smallest version of this. A street vendor sells two hundred plates of poha every morning. On Monday only a hundred and forty plates go out, and the vendor wants to know why. There are exactly two shapes the answer can take. Either the supplier turned up half empty and the vendor could not make the other sixty plates, or the office across the road sent its staff home for the day and sixty people who normally buy simply did not walk past. The count at the end of the morning is the same either way. The reasons are opposites.

A supply shock is the first of those two, scaled up. Something has cut what can be produced or delivered at any given price: a field under water, a refinery shut for repair, a shipping lane closed, a mine on strike, a well past its best years. The people who wanted the stuff still want exactly as much of it as they did the day before. No income changed, no plan changed, nobody decided to use less. A fall in demand pulls the quantity down as well, so the definition of a supply shock is entirely a statement about which side of the market moved and never a statement about how far the quantity fell. That is the trap. The quantity is the figure everybody sees first, and the quantity is the one figure that cannot separate the two.

Why do the price and the quantity end up going opposite ways?

Put the two lines back on the chart. Sellers part with more when the price is higher, so the supply line slopes up. Buyers take less when the price is higher, so the demand line slopes down. Where the two cross is the equilibriumThe single price at which the amount sellers want to sell and the amount buyers want to buy are the same number. On a chart it is simply where the two lines cross., and it fixes both the price and the quantity at the same moment.

Now take some of the supply away. At every price there is less on offer than yesterday, so the supply line slides bodily to the left. The demand line has not stirred, so at yesterday's price there are more buyers than barrels. The price has to climb until enough buyers have stepped back to match what is left, and the quantity that actually changes hands settles below where it was. Price up, quantity down.

Now do the mirror image. Take some of the demand away and leave supply exactly where it stood. At yesterday's price there are more barrels than buyers, so the price has to fall until the last barrels are no longer worth producing. Price down, quantity down. The quantity falls in both cases and only the price moves in opposite directions. The price is therefore the single observation that separates the two, and awkwardly the price is the one a headline about falling consumption most often leaves out.

There is a neat way to hold all of this. In the Sankhya illustration both lines are drawn equally steep, and with that assumption the arithmetic says something exact: the quantity responds to the sum of the two shifts and the price responds to the difference between them. A shift of 2.60 crore barrels a year on the supply side and a shift of 2.60 crore barrels a year on the demand side contribute the same amount to that sum, so both deliver the identical fall in quantity. The two shifts enter the difference with opposite signs, so the two shifts deliver opposite price moves. The equal steepness is a convenience of the drawing rather than a fact about any market, but the direction of the result survives without it.

Why can a single lever not put both halves right?

After a supply shock, anyone holding a lever that works on demand is in a genuinely awkward position. The policy rateThe rate a central bank sets for its own short lending to banks, which it moves to make borrowing across the economy dearer or cheaper. Covered in its own right under monetary policy. is the familiar such lever, and the reason for the awkwardness is worth being exact about.

The lever acts on one side only. Cool demand and both the price and the quantity go down together. Support demand and both go up together. A demand-side lever can do only that to a market: the lever moves the two readings in the same direction, every time, without exception.

Now look at what a supply shock left behind. The barrel price is above where it started and the barrels used are below where they started. The two readings sit on opposite sides of the base, so a lever that moves them together has to improve one and worsen the other. Cool demand hard enough to bring the Sankhya barrel price down from Rs 5,040/- to Rs 4,520/- and the barrels used fall further still, from 35.20 to 34.55 crore barrels a year. Support demand instead and the barrels used recover to 35.85 crore barrels a year while the price climbs further, to Rs 5,560/-. The lever faces a trade-off after a supply shock because the two things that went wrong went wrong in opposite directions, and that trade-off simply does not arise after a demand shock, where the price and the quantity fell together and one push in one direction lifts both.

The trade-off stands as stated, and no further. Which side of the trade-off to take depends on how long the shock is expected to run, what else is happening to prices, and which target the lever is meant to protect. All three change from one shock to the next, so no lever setting is the right one in general.

Why does the shortage sit there for so long?

A supply shock outlasts almost every answer aimed at it, and the reason is mechanical rather than political. Buyers respond to a price quickly. A household that finds fuel dearer this month drives a little less this month. New supply responds slowly. New supply is a physical object: a well has to be drilled, a plantation has to grow, a smelter has to be commissioned and staffed. In the Sankhya illustration demand answers a price within about one period and new supply answers within about four. Three periods are left over in which the gap between what is wanted and what exists simply stays open.

A supply shock lasts roughly as long as supply takes to rebuild. The length of a supply shock is a fact about drilling rigs and growing seasons rather than a verdict on anybody's competence. It also explains why the answer, when it finally arrives, tends to arrive late. By the time the new wells produce, buyers have already trimmed their use, so the fresh barrels land into a market that wants fewer of them than it did when the drilling was commissioned. The overshoot is the lag showing itself rather than a mistake by anyone in the market.

What did the same fall in barrels cost Sankhya from each cause?

The Republic of Sankhya uses 36.50 crore barrels of oil a year and pays Rs 4,000/- a barrel, for an import billThe rupee amount a country pays out for the goods it buys from abroad over a period. Covered in its own right under the external account. of Rs 1,46,000 crore a year. Now run the shock through it twice. Under the supply case the barrel price goes to Rs 5,040/- and use falls to 35.20 crore barrels a year, for a bill of Rs 1,77,408 crore. Under the demand case use falls to exactly the same 35.20 crore barrels a year, the barrel price goes to Rs 2,960/-, and the bill is Rs 1,04,192 crore. Sankhya bought 1.30 crore fewer barrels in both cases, and paid Rs 31,408 crore more in one and Rs 41,808 crore less in the other.

The first two rows are the entire lesson. Sit with the two rows for a second. A reader handed only the quantity row sees a fall of 3.56 per cent and has no way to tell which column produced it. A reader handed the price row knows immediately. Paying more for less is the signature of a supply shock in an import bill: the price effect and the volume effect pull opposite ways and the price effect wins comfortably.

The rise in the bill splits cleanly into those two effects, and it is worth doing the split because it holds a figure that appears in the wider Sankhya illustration. Keep the barrels at 36.50 crore and let only the price move: the bill would be 36.50 crore barrels at Rs 5,040/-, or Rs 1,83,960 crore, a rise of Rs 37,960 crore. The Rs 37,960 crore rise is the price effect on its own, and the price effect is the figure a reader will meet quoted as the cost of the oil move at unchanged volume. Then let the volume fall: 1.30 crore fewer barrels a year at Rs 5,040/- each takes Rs 6,552 crore back off, and Rs 1,83,960 crore less Rs 6,552 crore is the Rs 1,77,408 crore Sankhya actually pays. The Rs 1,83,960 crore figure is the bill at unchanged volume and the Rs 1,77,408 crore figure is the bill once the volume gives way, and both are rises on Rs 1,46,000 crore.

How can the two be told apart from outside?

The chart is rarely handed over. A published figure arrives instead, usually a quantity, and the work runs backwards from that figure. Three observations help, and none of the three closes the question by itself.

The first is the price over the same window. Use falling with the price rising points at the supply side; use falling with the price falling points at the demand side. The price is the strongest of the three observations and the one most often missing from the report in hand. The second is where the fall is concentrated. A broad fall in spending does not sort itself by which factory needed which raw material. So if the goods that lean hardest on the missing input fell hardest, the concentration points at the input rather than at spending in general. The third is whether a core measureA price index built with the most jumpy items taken out, so that a broad drift in prices can be seen without a single bad harvest or fuel move dominating it. Covered in its own right under inflation. moved as well as the headline one. A move confined to the volatile itemsThe parts of a price basket that jump around most from month to month, typically food and fuel, whose swings often reverse rather than persist. Covered in its own right under inflation. points at those items' own supply rather than at a general pull from demand.

None of the three observations settles the question alone, and a reader holding all three still has a judgement to make rather than an answer to read off. A price can move for a third reason entirely, including the currency a commodity is quoted in. The goods that lean on a disrupted input can also be the ones households cut first when money is tight. And an input large enough will work its way into a core measure too, given enough time. The three observations narrow the question from anybody's guess to a small set of things that can be gone and checked.

What does calling something a shock not establish?

The word does one job. The word says the change entered from a particular side, the seller's side rather than the buyer's, and the word says nothing else at all. Not the size: a supply shock can be a rounding error on a national account. Not the duration: a supply shock can clear within a fortnight when a port reopens, or sit for four years while a mine is rebuilt. Not whether it was foreseen: a monsoon that fails after two dry years is a supply shock that half the market was already positioned for, and it is no less a supply shock for having been widely expected.

The word shock names a direction of origin and nothing more. A note that uses the word as a synonym for large or sudden has said considerably less than it appears to have said. The three things the word does not carry are exactly the three a reader needs, and each of the three has a practical use. How large decides whether it is worth anybody's attention. How long decides whether a response has time to work. Whether it was expected decides how much of the move was already in the price before the news arrived.

What does an analyst actually split a price move into?

An analyst covering a company that buys a commodity, a lender setting a working capital limit against that same input, and a household deciding whether this month's fuel bill is the new normal are all doing the same work in different clothes. Only a move from the input side carries the four period tail behind it. Each of the three is trying to split an observed price move into the part that came from the input side and the part that came from spending. A price move driven by buyers coming back can reverse the moment buyers step away again. A price move driven by a well going offline stays put until another well comes online, and no amount of cooling demand shortens that queue.

The split matters because it decides which figure to forecast forward and which to treat as noise, and it is the reason two analysts holding the identical price series can write opposite notes without either of them making an arithmetic error. The practical version is short. Take the move. Ask what happened on the input side over the same window, using the three observations above. Attach the persistent part to the input and only the remainder to spending. Then say out loud how much of the answer rests on the split. A note that hides its own split is a note whose conclusion cannot be checked. The same discipline runs through pass-throughHow much of a change in one price ends up showing in another one further along the chain, such as a crude price reaching a pump price. Covered in its own right under inflation. work, where an input move gets traced along the chain and loses some of itself at every step.

The Republic of Sankhya and its partner Marut are invented.
Educational material. Not advice on any investment, tax, budget or market position.