I have done before and after power tests on K&Ns and similar filters, both on engine and chassis dynos, for magazine stories.
Whatever you get comes in at the upper end. Whether you get anything depends on how well or poorly the original filter flowed. Often whatever restriction there is come from the filter housing, not the filter element. I've seen as much as 5 hp on a medium performance, otherwise stock EFI V8 on a chassis dyno from a drop in element, but it's very hard to pull that number out of all the "noise" and variables in a typical chassis dyno run. Typically, you see nothing, or next to it, from an element change alone. On an engine dyno, you may see a bit more (because you can eliminate so many of the variables), but I've only run full intake systems, not drop in elements. Seen 15-20 hp once or twice back in the old carburetor days with a full system. No matter what, however, the gain is all at the upper end, wich isn't always useable day-to-day, and whether, or how much, you get will depend on how bad the OE system is versus how good the new on is. Many modern stock systems are pretty good, so the replacements have to be great to show a gain. In general, the only advantage I see to a replacement HP drop in element is that you can clean it yourself rather than replace it.
There's no fuel mileage gain to a free flow element on an EFI engine, though there may be a slight one on a carbureted one.
Don't forget that there are issue of filtering ability with the oil cotton gauze filters (K&Ns and everyone else's). Some of the dry elements using the latest whiz bang media are much better in that regard, though they down flow as much air.
