In a nutshell: depends on your starting point. If you've got relatively good OEM wheel cover aero you won't likely be able to detect the difference (which isn't the same thing as saying further smoothing don't work - it'll just be a small effect).
However if your OEM wheels are horrid air blenders, fixing the problem can be a significant improvement.
I have some numbers, not MPG improvement however. I have a Prius, and modified the trim rings with .010" thick aluminum flashing. This gives a flat hub cap. There are pictures on this site, and you can search them out.
Recently, I needed to do a tire rotation. So, I took off the trimring/hubcaps, and cleaned everything up in preparation to going to tire shop. While I had them off, I did a few coasts down a hill. At the base of the hill the road goes back to level. I would start off at 35 mph, go into neutral, and let the car coast up to 45 mph. I have a CANUSB hooked up to an EEEPC ASUS small laptop computer running MYCANSCAN. It gives speed in .1 mph units.
The mycanscan cycles about once a second. Without the trimring/hubcaps, the speed would peak and then fall off by .1 mpg or more the next cycle. With the hubcaps the speed would not drop to the next .1 mpg slower till after 3 or 4 cycles occured.
This year I am averaging 65 mpg plus on my commutes, versus 58 last year (12 percent improvement). The middle 2/3's of my commute is interstate highway. I attribute allot of the improvement to using Super Highway Mode. The hubcaps allow me to run SHM at 53 mph, and 1280 rpm. Versus more like 50 mph other people report for this mode. 53 mph is just practical on the interstate highway I (55 mph limit, lots of slow sections) commute on. 50 mph would not be. That is, if I did 50 mph, the gap to the illegal speed cars in front of me would open up to the point I would not catch up to these cars when they get to a slow portion of the road. Which would then be obstruction of traffic. I rarely get strong tail winds. The tail winds are during my morning commute, and the winds are calm in the morning. In the evening, however, winds can be strong and are are usually quartering winds (worse kind).
So, a guess is that the hubcaps are giving me 1/2 of that improvement, or about 3.5 mpg improvement. (58+ 3.5)/58 is a 6 percent improvement. Use that for any estimates on other vehicles that can reduce fuel consumption linearly (diesels, and or CVT cars). On cars with fixed gearing and stoiciometric burns, the improvement is probably going to be half that again, or 3 percent.
Thanks for that info, donee. That's the only effort I've seen anyone make so far at quantifying wheel disc effects. (The diesel/ecvt disclaimer is a nice touch.)
So what I am getting is they will either make no noticible improvment in gas mileage, or they will make some difference. If they make even a 1 or 2 mpg improvement in FE, then they would pay for themselves in 1 or 2 years.
Not sure if I want the kind that press on to the rims. I used to have pressed on hubcaps on my old 67 Mustang, and they were always falling off. I saw a screw on kind. Not sure what they screw on to though. Do holes need to be drilled into the rims or something?
I don't think anyone is going to be able to say with confidence whether it's going to make a difference or not on your particular car, with your current wheel covers as a starting point. (Unless they tried it on a car like yours, with those wheel covers.)
Why not try it out yourself? Cut out 4 cardboard discs, grab the duct tape and do some ABA coastdown tests / terminal velocity downhill runs to see if you can detect a difference?
"Racing disc worth it?" As part of a complete program to remove as much a possible the air's grasp on your vehicle, yes, unless you are driving in a vaccum, then don't worry about it,