Your input is great but it begs the question: why, then, isn't production being shut down during this time?
> we still make a small profit on low oil prices.
and
> when prices are low everybody expects to rise back to what we're used to again soon, because we're naturally optimistic. But that's not a law written in stone, it might be true that oil prices drop again in future, so we're better off selling it today.
So basically, a bird in the hand is worth two in the bush.
But once you've paid all that capital (and something like shale oil in the west has some of the highest relative capital costs) you're better to keep producing at anything north of opex, so maybe as low as $15-20 for a shale producer, or $5 for S.A.
This is all very estimated, but you start to see why they keep pumping.
Next add in things like transport and storage constraints, refinery capacity , etc and volumes range from zero to infinite regardless of prices.
source: we build modeling software for evaluation of these scenarios.
Keep in mind:
1) That the $20/bbl cost is likely with capex amortized on the assumption that the well isn't shut-in. You'd either have to add a one-time write-off for the capex balance (if you can afford to), or at least re-amortize. Either way, looking at the existing/assumed $/bbl cost vs market price isn't correct for most decision making purposes.
2) The act of shutting-in (and re-starting production) have direct costs to be accounted for. It's not as simple as pushing a button that activates a remote controlled valve. There's (almost) always physical work to be done, plus compliance / permitting work associated with any changes. More-so for restarts, if you assume the shut-in will be temporary, and want to account for that cost in decision making.
3) opex costs are largely still discrete to some extent (ie, not literally a cost in $/bbl, except pipeline transport and royalties), and can still be sunk costs.
Shutting-in wells doesn't necessarily reduce opex to $0, whether immediately or longer term.
Yearly maintenance might have been done last week. Tanker service already contracted for the year, or a direct pipeline already built. Lease and permit fees paid for the next X years. The lease might be contracted at some fixed cost plus royalties, for 10+ years, with the fixed cost guaranteed.
New for 2020: If the company took a PPP loan (or otherwise is trying to avoid layoffs) they have to keep surplus personnel, which then becomes a sunk cost.
Now if you look at exploration instead of production, it's much more likely to be a textbook case of if (cost > price), then (stop work immediately).
Things that are meant to keep moving tend not to fare well when they are left to sit for too long. If you have a classic car you have to drive it or everything starts to seize, separate, or gel.
I expect oil well equipment is not so different.