IP Becomes Part Of GDP Illustrating Problem With GDP
There are three possible way to measure output. One is to include the value of all transactions, the second is to include the value of only consumption and gross investment, the third is to include only consumption and net investment. GDP is essentially the second approach.
The problem with the first approach is that makes no distinction as to how much value was created in a certain transaction, creating the result that if a company outsourced one step in the production process rather than performing with its own employees then output would be considered higher even if the value of the final product wasn't bigger.
The second approach however also has problems as this revision is an example of. GDP will be about 3% higher even as the value of production isn't higher-after all, corporate profits won't be higher* and salary/wage income for workers certainly won't be higher as a result of this revision meaning that America won't really be richer despite the fact that its GDP will be 3% higher. This also creates problems when comparing U.S. GDP with GDP in other countries, unless they make the same methodological change. The reason why this problem has arisen is that it is not always obvious whether a certain corporate expenditure should be counted as "input cost" or "capital investment".
That is one of the reasons why I prefer the third approach. I usually use the GDP measure here simply because 1) Most other people use that measure, so to make myself understood I have to adapt 2) Usually the change in GDP is pretty much the same as the change in net domestic income so it makes little difference when analyzing economic trends. But net domestic income is nevertheless in principle the measure that makes most sense
*=at least real life corporate profits won't be higher. However, if the BEA fails to properly measure capital depreciation then its measure of profits might be higher.