The article said that Ford is offering goodwill payments to owners. Likely in an attempt to avoid a lawsuit.
In my opinion, the problem stems from the fuel economy ratings not actually being measures of fuel used. Unbeknownst to most people, the car is never actually driven on a track, and they never directly measure the fuel used.
http://www.fueleconomy.gov/feg/how_tested.shtml
The vehicle is accelerated, on the dynamometer, at very specific rates of acceleration, depending on the schedule being tested (city, highway, etc). This is how many of us believe the OEMs are achieving such good numbers (and why some modern powertrains feel so sluggish in certain driving situations, especially some with CVTs or 7- and 8-speed transmissions). They program fuel delivery rates and transmission shift schedules to support those very specific rates of acceleration in the EPA testing, and optimize the operation to achieve good results. It seems to no longer matter how the car drives in real life. The powertrains seem to be programmed to hit the best marks in the EPA testing.
But anyway, they measure hydrocarbons in the exhaust and run that through a number of mathematical formulas to estimate the fuel economy. If Ford is correct, an estimate of total vehicle load horsepower is apparently part of the "maths" involved, so an error in any of those variables would cause a shift in the results.
Having said all that, the vehicles we've owned tend to be pretty close to the old testing regime, the pre-2008 definitions. I have never gotten as poor as the rated fuel economy using the 2008+ numbers. The 2007 CR-V is rated at 22/24/28, and that's right about what we get with our 2008 model, despite the 2008 officially being rated 20/22/26. Same with our 2005 MDX: its window sticker showed 17/19/23, and that's right what we get; the "revised" estimate is 15/17/21, and we've never gotten as low as that.