Strategic Petroleum Reserves: Buffering an Oil Shock

A strategic petroleum reserve is crude oil held in physical storage so that a country can keep using oil for a while after its usual supply stops. The question a reserve answers is how long and not how much, so its size is reported in days of use rather than in money. A reserve buys time, and time is a narrower thing than a lower price.

A reserve rests on two ideas already built elsewhere. The first is a supply shockOutput or delivery drops while buyers still want exactly what they wanted before. Where one comes from, and how it travels through a market, is built separately., where the barrels a country normally receives stop arriving while every engine, generator and delivery van that burns them keeps running exactly as before. The second is the lag on the answer to that shock. New production takes about four periods to arrive, and demand answers in about one, so there is a stretch of calendar time in which the shortage exists and nothing anyone starts today has finished. A strategic petroleum reserve is a store built to sit inside that stretch of calendar time, and keeping the calendar in view explains most of what such a store can and cannot do.

All the arithmetic below belongs to the Republic of Sankhya, an invented country, so that a whole reserve position can be worked with every step exposed. Sankhya burns 36.5 crore barrels of crude a year, or exactly 0.1 crore barrels a day, and at the base rate a barrel costs Rs 4,000/-. The burn rate and the barrel price produce every other figure here, and both land the divisions on whole numbers rather than on decimals nobody trusts.

What is a strategic petroleum reserve, and what is it not?

A strategic petroleum reserve is a physical store of crude oil. Actual barrels, in actual tanks or underground caverns, held on behalf of a state and kept apart from the working stock that refinersPlants that take crude oil in and turn it into products such as diesel, petrol and cooking fuel. Each plant is built to run a particular range of crude and not any crude at all. hold to keep their plants running through an ordinary month. The distinction matters because it decides what the store can be asked to do. The store holds oil, so it can hand over oil. The store holds nothing else, so it can hand over nothing else.

A store answers an interruption, and an interruption is a different problem from an expense. Those two get confused constantly, and the confusion is not careless: both of them show up in the same place, which is a household or a country paying more for oil this month than last month. But the causes are unrelated. An interruption means the barrels are not there. An expense means the barrels are there and cost more. A store of barrels is a complete answer to the first and no answer at all to the second.

Picture a household that keeps one spare cooking gas cylinder in the corner of the kitchen. The spare answers exactly one situation: the delivery does not come on the day it was supposed to, and dinner still gets cooked. Now suppose instead that the price of a refill rises permanently. The spare cylinder in the corner does nothing whatsoever about that. The household is holding one cylinder of gas it will burn once and then have to buy again at the new price, and the spare has not made a single future refill cheaper. Nothing about a country holding 9 crore barrels changes that logic. The scale changes; the question the store can answer does not.

How is a reserve's size measured, and in what unit?

Sankhya holds 9 crore barrels. Whether 9 crore barrels is a lot depends entirely on how fast the country gets through barrels, so on its own that number says almost nothing. So the size is stated as a division. Sankhya uses 0.1 crore barrels a day. Nine crore barrels divided by 0.1 crore barrels a day gives 90.00, and because the thing divided by was a day, the answer is 90.00 days. Ninety days is the reserve cover, and the cover is the figure a reader is being handed whenever a reserve is described.

A currency reserve is measured in months of imports, and a petroleum reserve is measured in days of use, and a reader who carries the number across without carrying the unit has misread by a factor of thirty. Both are cover figures and both are built the same way, by dividing a stock by a rate of consumption. The denominator differs, and the denominator sets the unit. Foreign exchange cover sits on one month of goods importsThe value of physical products bought from abroad in a period. Goods imports are published in trade releases rather than in a reserve release, and how the figure is compiled is covered separately., so the answer comes out in months. Petroleum cover divides by one day of use, so the answer comes out in days. A 90 written down with no unit attached could mean three months or two hundred and twenty five years, and nothing in the digits settles which.

Working the same store the other way gives the money view. Nine crore barrels at the base price of Rs 4,000/- a barrel is Rs 36,000 crore of oil. Sankhya uses 0.1 crore barrels a day, and at that price one day of use is Rs 400 crore. Ninety days at Rs 400 crore a day is Rs 36,000 crore, so the two routes close on each other exactly. The match is not a cross-check. The same three numbers were used both times, and the second route is the first one rearranged. The match does show that the money figure and the day figure are the same fact, and only one of the two is stable.

Here is why days won. Suppose every barrel in the world became dearer overnight while nothing physical changed. The Sankhya store would be worth more in rupees, and it would still last exactly 90.00 days. The rupee value of a store moves with the price of the thing stored, so it reports the price rather than the store. Nothing moves the day count except the barrels or the burn rate, so the day count is the honest measure of a store. Notice also how small the store is against a year: 9 crore barrels against annual use of 36.5 crore barrels is 24.66 per cent of a year, and 90 days against 365 days is the same 24.66 per cent, as it must be.

What does WTI Crude Oil mean, and why does a store reference a grade?

West Texas Intermediate (WTI) Crude Oil is a real benchmark grade, and Brent is another one. A grade is not a number and carries no number inside it. A grade is a stated quality at a stated delivery point, so a grade is a definition rather than a price. The quality half is a written specification: how dense the crude is, how much sulphur contentThe proportion of sulphur in a crude oil. More of it means more processing to strip it out before the products can be sold, so plants care which crude they receive. it carries, and the other properties that decide what can be made out of it, each written as a range rather than as a single value. The delivery point half is a named place where barrels meeting that specification change hands. Put together, the two halves make a phrase that means one exact thing to everybody who uses it.

Naming one exact thing is what a benchmark grade is for. Without one, every single agreement to hand over oil would have to describe a barrel from scratch, listing every property and arguing about each one. With one, a futures contractA standardised agreement to hand over a stated quantity of a stated thing at a stated future time. The exchange writes the terms so that every contract of that name is identical. Covered separately. or a delivery agreement can say the grade and the quantity and stop there, and both sides know exactly what has been agreed. The same trick runs through every physical market that trades in bulk. A wheat contract that names a variety and a delivery yard has done the same job in the same way, and nobody writing it has to explain what wheat is.

Now the reason a store cares. Oil is not one substance. Two barrels can both be crude oil and be different enough that a plant built for one cannot process the other properly, or gets a very different mix of products out of it. So when a country puts barrels into storage, what goes in has a specification, and when barrels come out, what comes out has a specification. A release that hands a refiner barrels it cannot run has moved oil without solving anything. Naming a grade is how the store, the refiner and the agreement between them all say the same thing about the same barrels.

A price for West Texas Intermediate, a price for Brent, the difference between them, a level on any day: those are readings, they move, and they belong to the venues that publish them. The meaning of the word grade is what travels, and that meaning does not move.

What can a release from a reserve actually do?

A release does one mechanical thing. The store hands over barrels at the rate they are missing, day after day, until either the interruption ends or the store empties. Sankhya loses its usual supply, use stays at 0.1 crore barrels a day because nobody has told the buses to stop running, and the store hands over 0.1 crore barrels every day to fill the hole. Nothing clever is happening. Barrels are moving from a tank to a refinery instead of from a ship to a refinery.

The handover achieves one thing, a stretch of calendar. New supply takes four periods to arrive, and taking a period as 30 days, that is 120 days of waiting before the answer to the shortage shows up. The Sankhya store covers 90.00 days of it. So the store carries the country through three of the four periods, and 30 days at the end are not carried: 30 days at 0.1 crore barrels a day is 3 crore barrels that have to come from somewhere else, worth Rs 12,000 crore at the base price of Rs 4,000/- a barrel. Three periods out of four is the honest reading of this store against this lag, and it is worth stating plainly rather than rounding into a claim that the store covers the wait.

A release buys time and nothing else. Any effect on the price is a by-product of putting barrels into a market, not the purpose of putting them there. This is the sentence to hold on to, because it is the one that keeps every later question straightforward. Yes, barrels arriving in a market where barrels were scarce can change what the marginal buyer pays. But the release was not sized to a price, it was sized to a gap, and it stops when the barrels stop, whatever the price is doing on that day. Treating the price movement as the objective is how a store gets emptied against a problem it was never able to solve.

The household version again. A spare cylinder in the corner does not make the household richer and does not lower the price at the gas agency. The spare lets dinner get cooked on the days between the missed delivery and the next one. If the gap runs longer than one cylinder, the household is standing in the same queue as everyone else, just later.

What can a reserve not do, however large it is?

The barrels in a store are finite, they get handed over once, and then the store has to be filled again. Filling it again is a purchase in the spot marketThe market for buying something for delivery now rather than at an agreed future date. Whoever buys there pays whatever the market is asking on the day. like any other purchase, made at whatever a barrel costs on the days the refilling happens. Refilling at whatever the market asks settles most of what a reserve cannot do.

Run the round trip on the Sankhya store. Nine crore barrels leave, and at the base price of Rs 4,000/- a barrel that is Rs 36,000 crore of oil handed over. Now refill the same 9 crore barrels after the shock this whole sequence works on, where the rupee cost of a barrel has moved to Rs 5,040/-. Nine crore barrels at Rs 5,040/- is Rs 45,360 crore. Sankhya has put back exactly the same number of barrels and paid Rs 9,360 crore more to do it, or 26.00 per cent more for a physically identical store. Nothing was created anywhere in that loop.

A store moves consumption through time and does not create any. Every barrel used during the release is a barrel not available later, and every barrel put back later is a barrel bought out of the same market everyone else is buying from. Moving consumption through time is why a reserve cannot make oil lastingly cheaper. A store puts barrels into a market on the days it is releasing them, a real effect on those days, and takes barrels back out on the days it is refilling, the same effect with the sign reversed.

The harder version of the same point is what a store does against a permanent change in supply. If the barrels are gone for good, rather than gone for a stretch, then the store runs its 90.00 days and empties, and on day 91 the country is in exactly the position it would have been in on day 1, except that it now also has an empty store to fill. A store is built on the assumption that the interruption ends. When that assumption is wrong, the store does not fail gracefully. The store simply runs out, and the underlying problem is untouched.

What happens to the same store against two different interruptions?

Now put one store against two situations and change nothing about the store. Sankhya holds 9 crore barrels, uses 0.1 crore barrels a day, and therefore carries 90.00 days of cover in both runs below. Take a period as 30 days, so the two runs can be read side by side.

In the first run the interruption lasts three periods, or 90 days. The store hands over 0.1 crore barrels a day for 90 days, uses up all 9 crore barrels, and the supply comes back on the day the last barrel leaves. The country got through without use falling, and there is nothing left over. In the second run the interruption lasts six periods, or 180 days. The store behaves in exactly the same way and empties on day 90, and there are 90 days still to go. The remaining 90 days need 9 crore barrels that the store no longer has, worth Rs 36,000 crore at the base price, and those barrels have to be found somewhere else or the use has to fall.

The same store was sufficient in one run and exhausted in the other, and nothing about the store changed between them. Not the barrels, not the burn rate, not the cover figure. The length of the interruption changed, and the length of an interruption is a property of the world rather than a property of the store. A cover figure on its own can therefore never be read as a verdict, and the comparison rather than the figure is what carries the meaning.

What does a cover figure leave out?

A cover figure of 90.00 days is a division and nothing more, and three things sit outside it that decide whether those days ever get delivered. The first is speed. Cover assumes the barrels can leave the store as fast as they are needed, and a store that can physically move only 0.05 crore barrels a day cannot replace use of 0.1 crore barrels a day. Such a store would fill half the gap for 180 days instead of the whole gap for 90, and the arithmetic ceiling of 90.00 days would never be reached in the form the figure implies.

The second is grade. A store holds barrels of a stated quality, and the refineries that would receive them are built for particular qualities. If what is held is not what the receiving plants can run, some part of the store is not usable against this interruption even though it is fully counted in the cover figure. The third is the interruption itself. The figure knows how many days the store lasts. The figure knows nothing about how many days the shortage lasts, and that is the number it would have to be set against to mean anything.

A cover figure of 90.00 days is an arithmetic ceiling. The ceiling states what the barrels allow, not what the store will deliver. A cover figure carries no verdict. A cover figure carries a division. Everything that would turn it into a judgement lives outside it, and those things have to be gathered separately.

Where does the reading of a reserve go wrong?

What sits outside the store itself?