When the price falls, the same thing happens.
The price is market value, not what it cost them to fill the tank in the ground.
Else how do you price fuel in a 10K gal tank that is part at 2.00 / gal and part at 2.50 / gal ? Now how do you price it when that ratio ( just call it 1:1.2 ) is filled again with more expensive fuel. Bit of algebra ( it could be done ) but the price now fluctuates based upon how much of a given fuel is sold at a station and the day they get a delivery..
If the premium tank is filled the same day at 87, and the 87 tank is drained faster and filled with more expensive fuel, now you have premium being cheaper than 87.

Common sense shows that spot market pricing is the only way to have even fuel prices.

Price wise it does, as far as only the raw material costs delta.

DEC 2008, avg gas price ( depending on the state ) was 1.61 / gal with Brent @ 35
In JAN -2008 with brent @ 1.00 / brl, avg gas price was ~ 3.05 / gal with brent @ 100.00 + / brl.

65.00 * 0.025 = 1.625
JAN 3.05 - Delta 1.625 = Raw Material pricing of 1.425 / gal in DEC-2008.

go from DEC-2008 to today
1.61 + 2.20 = $ 3.81.
Google found prices in the 3.65 range for 25-FEB.

Was the 1.61 price in the same area ( or avg ) I found the 3.65 ( taxes )??

Was 1.61 price reflecting supply - demand pricing ?

Take Raw material pricing estimate of DEC-2008 @ 1.425 / gal + delta in raw material from DEC-08 to FEB-12 of 2.20 / gal = 3.625 / gal.

This indicates that the DEC-08 $1.61 price had a supply - demand delta built into it. The Raw material formula does not account for this pricing change.

Not sure what percentage you are talking about, but the math shows the formula is fairly accurate for what raw material costs does to the price of finished product.