My dissertation focuses on how ETFs affect the underlying assets. The first chapter studies how bond ETFs affect the liquidity of the underlying corporate bonds. The second chapter complements the first chapter by analyzing how to properly measure liquidity in the bond market. The third chapter studies how ETFs affect the liquidity of the underlying stocks. In chapter 1, "How do ETFs Affect the Liquidity of the Underlying Corporate Bonds?", I find that corporate bond ETFs improve the liquidity of the underlying bonds. I identify ETF arbitrage as an important channel for this positive impact. In the cross-section, the impact of ETFs is stronger on the less liquid bonds because these bonds are rarely traded by other types of investors. In chapter 2, "Are Corporate Bond Liquidity Measures Reliable?", I analyze corporate bond liquidity measures by benchmarking them against the effective spreads calculated using Bloomberg quotes. I find that most liquidity measures tend to underestimate transaction costs, and many do not correlate well with the effective spreads. Among all the measures, Roll's measure is the best at estimating the magnitudes of trading costs, and regression-based methods are best at identifying the cross-sectional variation in bond liquidity. I then propose a procedure to compute effective spreads for corporate bonds. In chapter 3, "How do ETFs Affect the Liquidity of the Underlying Stocks?", I find that ETF ownership is associated with higher transaction costs in the underlying stocks after 2010, but not from 2003 to 2010. The impact of ETF ownership is larger on large caps than small caps. Regarding the channels, I find that ETFs do not siphon trading away from the underlying market. On the other hand, ETF ownership causes liquidity to worsen during market stress for small stocks. My findings highlight that the impact of ETFs could be complicated, and differ for different periods and different stocks.