There is an idea that the Engine Management System can sense the higher octane and adjust the timing to eek out better MPGs. I have not been able to detect improved mileage using 91 over several test tanks. Now it's far more expensive I haven't revisited. There is now a more local place that has 0% ethanol super that I will try when I get a chance. No 93 available here in the West.
It's a little more complex than this.
Ethanol blends actually increase the octane rating of the fuel. This allows the engine to add more timing, but because ethanol contains less chemical energy than the gasoline blend, more fuel must also be added to maintain the proper A/F (adjusted) ratio and keep from bouncing off the knock sensor.
A "flex fuel" vehicle for instance, will get WORSE gas mileage when running E85 than when running E15, however, E85 costs less, so it becomes a complex exercise in math/economics.
Generally speaking, E85 contains about 30% less energy than typical E10 blends, but also costs less (the savings varies wildly depending on availability in your area.)
Compare and contrast this to the issue of 93 vs. 87 (which are both still blended with ethanol in most places across the country) and it gets a little less complex, but bottom line, GENERALLY speaking, higher octane usually just equates to smoother running and the mileage/cost gets washed away (because higher octane usually contains less actual energy).
On my Ford, I have an NGauge computer and I can actually monitor every single signal the car produces. So on my screen I typically show A/F ratio, Spark Advance, Knock Reduction (the amount of timing the engine is pulling out for knock), cylinder head temps, etc.
When running E85, I can produce a good bit more power, but my fuel mileage goes way down (but E-85 is a good bit cheaper than premium).
