The basic problem is that economics is a closed system. All we can do is model it. Those models are necessarily imperfect. The question then becomes what imperfections do you tolerate and what imperfections cause you to change the model?

GDP wasn't the leading economic indicator when I was growing up. It was GNP, which measures productivity of a nation's citizens no matter where they are (as opposed to GDP, which measures productivity of anyone no matter their citizenship within a nation's borders). For whatever reason, the US flipped from GNP to GDP in 1991 and in doing so, altered the way the economy was measured.

The emphasis on GDP is consistent with measmement considerations. Data from BEA’s direct investment survey, which is one of the primary sources for estimating factor income payments and receipts, are not avail- able for the first two of the three quarterly estimates of GNP. For these two estimates, factor income payments and receipts are based on judgments about trends in the pace of economic activity in the United States and abroad and about the value of the dollar in foreign countries, on announced profits of individual companies, and on other information. Even when all of the source data become available, BEA does not have the information needed to make a Ml set of adjustments to reflect the concepts underlying the NIPA’s. For example, the profits of foreign affiliates do not include inventory valuation and capital consumption adjustments, and they are affected by intracompany transfer prices and exchange rates.

PDF

Straight out of the horse's mouth - that's the Bureau of Economic Analysis, saying they're switching from one primary measurement to another - better compatibility with Europe, more data less forecasting, etc. In doing so, however, they fundamentally changed the tools used to shape policy.

That's the argument here: GDP does not reflect the rules on the ground accurately enough to dictate policy when it comes to government spending. Whether an action is "intermediary" or not is irrelevant - economics is, again, a closed system and pointing out an externality does not invalidate the externality, in invalidates the model.

Such as the personal-private spending thing. By pointing out that human capital is the primary economic asset of the US economy, and pointing out that said capital is refined via public school, and then illustrating that the value of the schooling is calculated based on what it cost vs. what it earned, the author is demonstrating a weakness in the model that is not factored into GDP.

I'm most of the way through Piketty at this point. He made a couple of interesting points in the bit I read yesterday:

1) Between 10 and 20% of the world's economy is socked away where we can't see it in offshore accounts

2) The stock market crash of 1929 became the Great Depression because economic policy at the time was "survival of the fittest;" the Great Recession didn't become the Great Depression II because economic policy at the time was "liquidity will save our asses."

The former statistic illustrates that our modeling and our measurement needs to be better. The second statistic illustrates the benefits of better modeling and measurement. Imagine what our social policy would be if we were measuring unemployment accurately?