I'm not sure if this makes any sense to an economist, but think of it this way:
If you found out the price of gas was GOING to go up, what would you do? Fill up your tank. When the price of gas goes up, you still have a full tank purchased at the cheap price.
Now say you know the price of gas is going to go down. You hold off refilling until the price drops.
This would mean that you are buying before a price increase, and after a price drop. Thus, the gas station would lose money whenever the price of gas is volatile.
"But I don't do that," you're going to say, "I don't know when the price of gas is *going* to go up."
Precisely. The gas stations raise the price of gas when their costs WILL go up, not once they HAVE gone up, preventing drivers from profiting from price volatility.
The fact that this corresponds with the lag I described earlier works out well economically. If these effects did not cancel each other out so cleanly, perhaps you would see other mechanisms for balancing them better, like cars with 30-gallon gas tanks.
In larger-scale sales without such a balancing effect, the system has had to create to product suppliers from losing money due to volatility. Refineries will contract for millions of gallons and price volatility could be devestating to their suppliers. Picture storing a million gallons of gas right before a price hike, or allowing a million-gallon tank to run almost dry while you wait for a price break! Futures are sold precisely as an insurance against price volatility for suppliers.
That makes a certain amount of sense, except that if all gas stations are constrained to similar extent by the actual cost of gas on the downward side, then one could make money by NOT jacking the price on the upside – at least not as greatly as the other stations – and pull in volume sales at a slightly increased margin of profit. This competition, at least in theory, should push stations to keep prices consistent with actual gas costs.