Yeah, first of all, 3% in 14 years is non-trivial, so this very graph shows that it would make more sense to include food and energy in the CPI, especially if the argument to not include it is "it makes CPI less predictable". If you want predictable results, just calculate them and post them, you don't have to mangle real-world data. Specifically, I see a bump in 2008/2009 - do you think that people who had to buy food then were happy that CPI is predictable, when their bread was suddenly x% more expensive? And second thing is: this is a random graph from an institution that has a serious conflict of interest here, trying to show that their CPI is good. Why would they show anything else? The good idea would be to take a basket of commonly bought goods from 2000 and compare to the same basket in 2014, and check the prices. Simple things, you know: bread, milk, butter, gasoline, electricity, poultry, things like that.
As far as the purchasing power goes, I don't really know the definition you are using, or if Mict is using the same as you, and I don't really know any definition at all; so I can't tell whether he is right or not.
I'm also very interested in any definition that puts disposable income as something not containing food, but containing other stuff. I've always imagined that the food is the first thing to buy in any reasonable case, but maybe some people prefer to die of starvation under the roof, than of exposure with full stomach, so maybe some kind of payment for shelter?