Chapter 1 The rise of exchange-traded funds (ETFs) has attracted significant attention from investors, regulators, and academics alike. Despite the widely asserted importance of authorized participants (APs) to ETFs' structure and function, little is known about the effects of AP behavior on ETFs and ETFs' underlying assets. Using AP-identified regulatory data from FINRA, I show that APs' inventories do not affect ETF prices, contrary to past inventory literature. I find that when APs are inventory-neutral, liquidity shocks to the ETF do not transmit to the underlying assets. These results are largely unchanged in a high-volatility subsample, although there is some concern for thinly traded ETFs. In a daily panel, I show that APs' inventories are associated with their decisions to create or redeem ETF shares. These results suggest that the unique features of APs may benefit ETF market participants, especially in more liquid ETFs. Chapter 2 Exchange-traded funds (ETFs) trade throughout the day, just like their underlying securities. Yet the simple problem of knowing an ETF’s underlying portfolio value during the day has been elusive in publicly available data. I develop a new regression- based method to compute the ex-post underlying portfolio value of a US passive equity ETF on an intraday basis. Using widely available data, this method improves in three ways over the most popular method to date (intraday indicative values): (1) Higher time resolution, (2) No stale prices, and (3) Visibility of the underlying bid-ask spread. I also provide a LASSO version of the method, and I validate the results using official portfolio data from the DTCC, a large central clearinghouse for ETFs and their underlying securities. Chapter 3 Leveraged exchange-traded funds (Leveraged ETFs) advertise the ability to replicate the daily return of a given index on a levered or inverse basis. These products are highly risky, and the costs of trading leveraged ETFs are quite substantial, raising questions about which types of traders would choose to use them. I use confidential regulatory data to identify institutional and retail order flow in ETFs, and I demonstrate that institutional order flow to leveraged ETFs is uninformed.