If an ETF shareholder passively holds the shares, there is still active trading in the shares in the highly liquid basket by the AP's (to track the index), right? But doesn't this incur a cost to the passive shareholder? The AP's trade on risk free arbitrage, more so in volatile markets, does that trading incur a cost to passive ETF shareholders? Or does that cost (the other end to the risk free arbitrage is a cost to someone) only apply to the shareholders that trade the ETF shares (i.e. not passive)? Where do the costs of risk-free ETF arbitrage end up?
The following may seem somewhat biased but it is simply based on the fact that there is a risk free source of gain, so a rational agent would seek to exploit it:
The impression generally given is that only the next ETF shareholder pays a small "fee" to the previous shareholder they bought it from. However, I think this is not the case. I think it is an ongoing cost incurred on market prices (daily arbitraging), this is no free lunch, these costs are paid. And I think they eat away at (are transferred to) the general share value of all stocks, despite the apparent overall inflation, which is by and large central bank induced (as well as by mandatory pension inflows). I think this is significant.
Consider an AP and maximise these risk free profits: Imagine you have a market where there is one company and its bank, and the bank is also the AP. The company's stock goes in an ETF and there is one passive shareholder. The AP can do whatever. What would the AP do?
While it is known that there is information between AP's knowledge of next day data and today's trading for example, this question is about when and where the costs are or could be incurred.