In the FT, Greg Meyer ponders a puzzle: “A mystery is confounding the US oil market: when inventories rise, prices rise, too.”
Yes, it is normally the case that inventories and prices, and inventories and the spot-deferred spread, move in opposite directions. But this does not have to be the case.
The typical case is based on the following economic logic. Inventories respond mainly to current, and temporary, supply and demand shocks. If current demand falls, and this demand shock is anticipated to be temporary, then current availability rises relative to expected future availability. The efficient way to respond to this is to store more today because the commodity is abundant today relative to what is expected in the future, and efficient allocations move resources from where they are relatively abundant to where they are relatively scarce. Storage increases expected future availability, which depresses expected future prices. The nearby price must fall relative to the expected future price in order to encourage storage, and together the fall in the expected future price and the fall in the nearby price relative to the expected future price causes the nearby price to fall.
A similar story holds with respect to a temporary increase in current supply.
Parenthetically, the temporary nature of the shock is important in driving the change in storage because this causes a change in relative availability that is necessary to make it optimal to store more. A shock that is anticipated to persist does not change current availability relative to expected future availability, so there is no benefit to shifting resources through time via storage. A persistent shock causes a parallel shift (roughly) in the forward curve, and no change in storage. In my academic research, I show that in a dynamic storage model supply/demand shocks with a very short half-life (on the order of 30 days) drive storage behavior, and that very persistent shocks drive the overall level of prices.
But there are other kinds of shocks. One kind of shock is to anticipated future demand or supply. Let’s say supply is expected to decline in the future. This increase in expected future scarcity can be mitigated by storing more today (i.e., reducing current consumption). This spreads the effect of the anticipated future supply loss over time, and thereby smooths consumption in an efficient way. The only way to reduce current consumption in order to increase inventories is to increase the spot price. So in this scenario, (a) inventories and prices move in the same direction, and (b) inventories and calendar spread (deferred minus nearby) move in opposite directions in order to reward the higher amount of storage.
Here’s a real world example. The Energy Policy Act of 2005 mandated increased use of renewable fuels–notably ethanol–in future years. This caused an increase in anticipated future demand for corn used to produce ethanol. When the act was passed, the supply of corn was basically fixed. One way of responding to the expected increase in future corn demand was to store more immediately (thereby carrying current supplies into the future when demand was going to be higher). Given the fixed supply, the only way to achieve this higher storage (and hence reduced current consumption) was for prices to rise.
Therefore, one explanation for the positive co-movement between prices and inventories is a shock to the expected future supply/demand balance. For example, an increased likelihood that OPEC will extend its supply cuts beyond April could produce this result.
Another kind of shock that can lead to a positive co-movement between spot prices and inventories is a shock to supply/demand volatility: I discussed this in an early blog post, and later analyzed this formally in my 2011 book. (A good example of the synergy between blogging and rigorous research, BTW.)
The intuition is this. Inventories are a way of insuring against uncertainty: putting something aside for a rainy day, as it were. If fundamental economic uncertainty goes up, it is efficient to hold more inventory. Since supply is fixed in the short run, the only way to increase inventory is to reduce current consumption. The only way to increase current consumption is for spot prices to rise. Moreover, to compensate increased inventory holding, futures prices must rise relative to spot prices. Therefore, for this kind of shock (like a shock to future demand) the forward curve rises and becomes steeper (i.e., increased contango).
So although the positive co-movement between spot prices and inventory may be unusual, it can occur in a rational, efficient market. It depends on the underlying driving shock. The typical case occurs when shocks to current supply/demand dominate. The more unusual case occurs when the shocks are to expected future supply and demand, or to fundamental volatility.
This relates directly to something I mentioned in the “kill the economists” post yesterday. Specifically: never argue from a price change. It is necessary to understand what is causing the price change. When there are multiple shocks that can affect prices (e.g., supply and demand shocks; current or future shocks; shocks to supply/demand volatility as well as to the level of supply/demand), just looking at the pice movement is not sufficient to draw conclusions about either its effect, or its cause. Indeed, it is even misleading to talk about the “effect” of the price change, because the price change is itself the endogenous effect of underlying causes/shocks.
The usual way to sort out what is going on is to look at quantities as well as prices. For instance, in a simple supply-demand model if you see prices go down, that could be because supply rose or demand fell. You can figure out which only by observing quantity: if you see quantity fall, for instance, you know that a demand decline caused the movements.
This means that the recent co-movements in oil inventories and prices reflects market participants’ assessment that the supply/demand balance is expected to tighten in the future, or that fundamental uncertainty is going up, or both.