1. EIBT margins: "like for like" (i.e. assuming same industries)
2. EBIT margins: industry substitution
3. Net margins: interest cost
4. Net margins: tax
...
Its only the last part in this breakdown - industry substitution (2) - where i believe the margin COULD be more sticky.

It's certainly a factor. There seems to be quite strong disagreement among observers as to how large an effect it is. Even among thoughtful and statistically well informed ones.

On the one hand, Mr Hussman's chart where the ONLY difference between the red and blue lines is interest and tax makes a pretty compelling case that the widely fluctuating gap between the two is, well, interest and tax. It may not explain everything, but you can see the "tax and interest" gap narrowing a huge amount, and there is no other factor at work. There is also another 2022 study by the Fed, showing that 1/3 of all US profit growth in the prior 20 years was driven by falling tax and interest expenses.
federalreserve.gov - The coming long run slowdown in corporate profit growth and stock returns

On the other hand, a weirdly bullish note from GMO in mid 2020 says that the changing mix of companies is the single biggest factor. Particularly surprising as their reputation has become that of permabears.
Why We Are Not Worried About Elevated Profit Margins
(you don't have to register at their site, just search on the title and you'll see a PDF result).
Among other things, they note that there are two separate "change of mix" effects: low margin companies no longer in the index being replaced by higher margin new entrants, and faster-than-index growth of some large high margin businesses.

Personally I am no more likely to invest in the largest company in the index than in the 500th largest, so I tend to track the median number rather than the aggregate. This means the issue of how lasting the super-high margins are for a few super-big firms is not my prime concern, as it would be (definitely should be) for a cap weight index investor. The aggregate is dominated by a few large firms that one can't come up with a sense of cycle or direction any better than you can come up with a prediction of those few firms, so I stick to the average and median metrics. Pretty much by definition, there aren't very many firms that are atypical.

Jim