Guys, let me clue you in to some tech here. The stock intake tubing has hemholtz resonators, commonly called air silencers. What these really do is quite interesting. They affect the resonance wave tuning of the intake, and therefore the torque curve. If you think of airflow as a surfer catching a wave, this will make a little sense. The first wave goes into the engine. Any time there is a change in tubing size or shape, the wave generally turns around and goes back. The wave will go into the motor until the intake valve closes.
At that point, the wave bounces off the intake and runs back up towards the air filter. If this wave comes out the air filter, you'll hear it. It will be "intake noise". The worlds best example of intake noise is a ricer kid's Honda Civic with a K&N filter and a straight cold air kit. All the noise those things make isn't exhaust noise, it's intake noise!
To prevent this noise, a hemholtz resonator is installed in a strategic location, generally near the air filter. The pressure wave spreads into the resonator, where it is basically turned around and sent back up into the intake tubing towards the motor. Since the wave didn't exit backwards out the air filter, the noise didn't come out. So we can see how it is an air silencer after all.
Less known is the other reason why the resonators are there. Just like its faster for a surfer to ride a wave towards the beach, it's also faster for an air intake charge to ride a resonance wave into a motor. It has almost a supercharging effect, although very slight. Camshafts and aftermarket intake manifolds do a very similar thing. In racing, we focus these waves to hit at high rpms, where we're trying to make power.
The OEM focuses these waves at lower RPMs, in order to make torque right where you'll be driving most of the time. This helps mileage and low end torque. So does a smooth cold air intake system improve power? Yes, by ultimately unrestricting the factory air intake path. But it will get noisier due to lack of a hemholtz resonator. And noise restrictions is part of the OEM goal. It is also likely that you'll LOSE 5 - 15 ft lbs of torque at lower rpms, unless the intake tubing was so restrictive from the factory that it actually gains everywhere.
Another major misunderstood concept is the MAF sensor. I cannot overstate the importance of the MAF sensor and it's placement. With older 1 piece tube/sensor MAFs, changing the piping around the MAF is not always a killer in the tune, but it should be compensated for via dyno tuning the MAF Xfer function in the PCM calibration. With the newer slot-style MAFs (sensor separate from the tube), the tube itself acts as the MAF. The MAF sensor itself is measuring a percentage of a fixed tube size.
Rough example: You have a MAF that has a 1/2" sample tube, and the MAF tube is 4", then the MAF is measuring 1/8 of the total airflow thru the MAF tube. So if the sample tube detects 10 cfm, then the computer knows that the total cfm must be 80. So what happens (in these slot style MAFs) if we keep that same 1/2" sample tube, but install it in a 5" intake pipe? We're now measuring 1/10 of the total airflow, not 1/8. So the computer will see 8 cfm instead of 10 cfm. So it will put in 8x fuel instead of 10x fuel. You've just leaned your motor out by 20% across the board.
With such a drastic change, why then do so many guys do it and claim they have no problems? O2 sensors. They have a 25% authority in either direction (rich or lean) to fix problems like this. So they re-learn over time. So what's wrong with that, you ask? The computer ignores the O2s at WOT, and in open loop, and your motor runs 20% lean. This is a BIG problem with slot-style meters, which generally showed up around 2005 or so.
On older MAFs, a lesser problem is encountered, where the new cold air piping has different bends in it, and flows some air at an angle towards the sample tube, making it's readings skewed slightly, maybe 10% or less. Still a problem, but not nearly as bad. In either case, the correct fix is to dyno tune the motor on a load-bearing dyno and remap the airflow xfer function in the PCM.
Generally, gas mileage, power and engine safety will all be affected. And by the way, the MAF sensor determines engine load via PCM calculations. This load determines spark advance settings. So an artificially low load can and usually will cause an artificially high spark advance timing. This can cause detonation, and possibly engine failure. Moral of the story: If you change anything about EFI, be prepared to retune for it.