It is known that: When the price of an ETF is lower than its NAV (i.e. Net Asset Value, the value of underlying securities, per ETF share), the Authorized Participants essentially buy ETF shares from the market, then they exchange with the ETF company ETF shares with the shares of the underlying securities. This is an "in-kind" transaction, so nobody pays taxes. At this step, usually the fund company provides to the Authorized Participants the batch of shares that have the lowest cost basis (i.e. the shares that were bought at the lowest price), which are the ones that would generate more capital gain taxes. Then the Authorized Participants sell the securities on the market, get the cash and make a risk free profit (arbitrage) based the difference ETF price and NAV. Selling the securities on the market tends to lower the ETF NAV, buying the ETF on the market tends to increase ETF price. The process continues until the difference between ETF price and NAV is so small that makes this process not profitable anymore for the Authorized Participants. This process is necessary to align the ETF price to the NAV (often called "creation and redemption" mechanism of the ETF's)
Question: In the scenario above, the Authorized Participants sell the securities with low cost basis on the market, therefore they end up paying taxes on that transaction. Clearly there is a limit to the amount of taxes the Authorized Participants are willing to pay. If the taxes are too high, the taxes may make the arbitrage mechanism unprofitable for the Authorized Participants. So what are the "rules of the game" between the fund company and the Authorized Participants in this regard? Maybe the Authorized participants must always accept any cost basis from the fund company even if this would be unprofitable for them?