Economic statistics are used for different purposes. For example, there is often confusion over the Fed's use of price indices for predicting future inflation and the use of the indices to measure the cost of living. Those are different questions so it shouldn't surprise us that different measure of the price level may be needed to answer them. Surprisingly, the best measure of today's cost of living may not be the best prediction of the cost of living in the future, some measure of "core prices" may be better.
The lesson is that the variable that best measures economic forces at work in the economy may not be the best measure of other things that we care about (another example is the ex-ante real interest rate which measures the economic forces that affect decisions, and the ex-post real interest rate which measures the actual outcome, though this is a bit different than the other examples since it's about potentially using the wrong measure rather than mis-measuring a particular aggregate of interest).
In our national income accounts, measures such as GDP and GNP measure national income and other domestic variables, and those measures represent things we care about. This is not a well-formed thought, but I've been wondering if variables such as these are the best measures of the economic forces that exist in a global economy. This is not a novel thought, it is what's behind attempts to measure the "global output gap" as discussed recently by Federal Reserve Chairman Ben Bernanke, but I think it's a concept worth thinking through further.
If our national aggregates, things like the price level, the unemployment rate, and national income fail to fully capture the underlying economic forces operating on the domestic economy, if they are inconsistent with the underlying markets, then economic policy based upon them will deteriorate accordingly. There was a time when most of economic forces hitting the U.S. economy were the result of domestic influences so the aggregates served a dual purpose, but today the rest of the world is much more important. I am not trying to say that this makes our current measures useless for policy, they may still do very well, only that it's worth thinking about how best to capture the economic forces at play in the models we use to make policy decisions.