> If company A is worth 1 billion and company B is worth 2 billion yet they have equal investments in both, then they sell some of A to buy some of B.
This is a paradox -- both of these are examples of market movements which require the percentage of stocks held outside of index funds be equal to or greater than the volume required for the price movement. The critical point is much lower, and as more people buy into index funds, the active traders make more money rebalancing and therefore are able to increase their own positions, providing negative feedback and preventing the index fund ownership from rising more. I don't know what the asymptote is, but I suspect it's closer to or equal to 50%.
Another way to think of it is that markets would be incredibly illiquid with 99% index fund ownership -- that's not much ownership for market makers, but their profits would be tremendous due to the thin order books, driving in more market makers and thereby increasing their market share until it's no longer easy money.
My point about ownership percentage was simply that passive traders and active traders only interact when the total amount of stock changes. But, passive traders are simply speeking a fixed ownership percentage which means they don't change the final price only the active traders who must end up with that same 1% based on price movements.
If they own 1M$ of A and 1M$ of B, but B is worth more than A then they sell some of A to buy some of B.
Keep in mind that this doesn't (at least not naïvely) happen when one of the two increases or decreases in value relative to the other.
Also of note index funds do have feedback loops as they try to retain a balanced portfolio. If company A is worth 1 billion and company B is worth 2 billion yet they have equal dollar investments in both, then they sell some of A to buy some of B. This greatly magnifies the impact of active traders in a 99% ownership situation.