The change is much smaller than most people suppose. Large US companies have always been exporters, even 50+ years ago. IBM, anyone? For example 27% in 1992 vs 29% in 2019 is not exactly a night and day difference.

It was just a passing thought, not something I investigated at all.

But if you look at how GDP is calculated - GDP = Consumer + Govt + Inv + (Exp - Imp) then if we have a trade deficit, that subtracts from GDP. Furthermore, my point (I think) is that if a company builds a bunch of stuff in Asia, and sells much of that stuff in Asia, then none of that gets accounted for in GDP, yet the profits from that activity [eventually] flow to the company in the USA.

So, while IBM (and all sorts of other companies) exported a bunch 40-50 years ago, they didn't make all that much stuff outside the USA to be sold outside the USA. That is in fact the lament against globalization, that it hollowed out the USA's manufacturing sector.

Meanwhile, there is an offsetting change...revenues of non-US firms firms inside the US.

Great point! All those non-USA firms contribute to GDP, but they are listed on an external stock market (ignoring ADRs for now), and the profits flow to there.

It's an interesting discussion. It would be good if we could at least see the order of magnitude of the numbers. If 50 years ago, 5% of total US stock market earnings came from overseas manufacturing (someone else's GDP), and today 20% of total US stock market earnings come from overseas manufacturing (someone else's GDP), then we could adjust the numbers accordingly.