Because they are ill-defined conceptually, many official economic statistics fail to capture what they purport to measure. Figures on "poverty," for instance, are notorious in this regard. Is poverty an absolute or a relative condition? If the latter, what is the proper standard of comparison?
Citing GDP, CPI, or any other host of acronymious jargon to explain a fundamentally complex subject is an exercise in futility. Each of these measures is limited by scope: there is simply no way that anyone can perfectly compile all the relevant data to make an accurate measurement of whatever it is that is intended to be measured.
But even if economists could have perfect information, the methodology behind the various metrics is suspect. GDP, for example, attempts to account for government spending, even though it is generally accepted to be inherently inefficient. CPI ignores fuel and some food costs, even though these are primary expenditures for a good portion of US households. Poverty measures are often relative.
That doesn’t make these metrics necessarily useless. I lied. Most poverty measures are worthless, as is CPI. Seriously, what value is a price metric that doesn’t measure the most important prices on the grounds that they’re “too volatile?” Know why food and fuel prices are so volatile? It’s because of inflation, you dipsticks!
GDP is useful, but only as a comparative measure, and only for each given year (e.g. 2009 French GDP can be fairly compared to 2009 American GDP, but not 1950 French GDP). Frankly, economics is best viewed as a more of a philosophy than a science. There isn’t really any empiricism (in the sense of experimentation) that accompanies the dismal science, and most studies are based on past data, meaning that studies resemble history more than, say, physics.
I’m not saying that economics doesn’t have value. I’m just saying that most economic metrics are worthless.