I believe the composition of American business is an important factor when analyzing historical margins and what may be normal or elevated. ...
I have no clue where margins or returns on capital will go from here but I think the above are important considerations for anyone brave enough to put forward a forecast. ..

I think those factors are a whole lot more important when you're considering any one company, or a few very large ones, or even an industry. When you're considering the entire US economy as a whole, top level macro limits come into play. Not every student can be above average.

The aggregate revenue of all companies in the economy grows (pretty much axiomatically) at the same rate as GDP, which is slow. That total gets divided up into only so many sectors: some goes to corporate profits, some goes to taxes, some goes to labour, some as interest to private lenders, and so on. In aggregate, none of those pieces can grow faster than GDP except at the expense of one of the other pieces.

It follows that the appearance of a few absurdly profitable large companies with very high net profit margins more or less requires that some other part of the economy be doing worse: the profitability of those firms does not remove the fact that over time the pie doesn't grow faster than GDP. Either a bunch of other companies are doing that much worse in terms of net margins, and/or the profits are arriving at the expense of labour or taxes or private lenders. Which might be temporary. In short, a few wildly successful firms with high net margins can not explain why economy-wide net profits can be growing faster than GDP. Because of the necessity that the allocations add to 100%, the aggregate high net margins have to be explained by shrinking shares going to the other sectors. Taxes, private interest, labour.

What is that analysis missing?
Cross border effects are large for other countries, but the changes are not very material by comparison for an economy as large as the US, so ignoring those can still give meaningful results. The government deficit is a very complicating factor...a huge debt-powered fiscal stimulus (whether government or private) is great for company profits for a while, but Ricardo would tell us that it's only borrowed from the future and will have a compensating drag at some time in the future when the debt is reduced. You have to be careful not to assume that (say) the S&P 500 typifies all companies in the US both public and private--the public companies could pull ahead (or fall behind) relative to private companies for a while. Though that's probably not a big factor, since public companies include almost all the big ones.

Jim