With the price of oil having plummeted, there has been considerable discussion in the media and elsewhere about what will happen to the quantity supplied (pedantic note for colleagues and students: quantity supplied, NOT supply) of oil in the short run and in the long run.
This article in Slate has it almost right. The average costs per barrel of oil include considerable sunk costs that are unrecoverable. If the price of oil is expected to stay below these expected long-run average costs, then no more new wells will be started in those oil fields.
In general, if the price of oil is expected to remain below $65/bbl, then there won't likely be many new shale oil facilities that will make it beyond the planning stage. And if the price of oil is expected to remain down nearer to $50/bbl for a long period, no new projects are likely to be begun in the arctic, tar sands, or deep sea [graph from Slate link]:
But for the short run, very little if any of the existing wells will be shut down. Here's the Econ 100 explanation:
The cost of drilling those wells has already occurred. It is a sunk cost (both in jargon and literally, I guess). The only relevant decision for an oil company is whether the revenue from continued pumping will offset the extra costs of doing the pumping, transportation, and marketing. These extra costs are generally referred to by economists as marginal costs, but what we mean is "extra" or "incremental" costs. Accountants sometimes use the term "direct costs" to refer to approximately the same thing.
What are the extra costs of pumping an additional barrel of oil, and how do they compare with the extra revenue the firm gets from selling that barrel? In economics jargon, pump another barrel so long as the MR>MC. In the case of oil pumping firms, MR=P probably.
Here is a graph (also from the Slate link) of the marginal costs of pumping and selling oil from various sources. The graph labels these as "cash costs" but it means marginal costs:
This graph makes the point really well. So long as the oil companies are receiving enough to cover these marginal costs, they will keep pumping the oil. And that will occur so long as the spot price of oil exceeds about $40/bbl. Pumping oil at those prices will cover the variable costs of pumping andmake some contribution toward covering some of the overhead/fixed/sunk costs.
There is an exception not addressed in the article, however. If the costs of stopping and starting the pumping process are low, some oil companies may choose to stop pumping if they expect oil prices to rise in the future.
In this case, the marginal cost of pumping oil now is not just the extraction, transportation, and marketing cost; it is also the present value of lost higher revenues in the future, which of course depend on the expectations people in each oil company have about future prices for oil. If they expect prices to rebound in the near future, they may want to curtail some pumping; if they expect prices to remain low for the foreseeable future, they may decide to keep pumping.
Note, though, that this decision depends only on their expectations about future prices of oil and has very little to do with the marginal costs of pumping. Or, to put it differently, the marginal opportunity costs of selling oil for cheap now are the possible foregone revenues from waiting.