We examine liquidity creation per unit of assets by banks subject to the Liquidity Coverage Ratio (LCR) using the liquidity measures Liquidity Mismatch Index (LMI) (Bai et al., 2018) and BB (Berger and Bouwman, 2009). We identify the LCR effects through time and cross-section effects, specific LCR-constrained balance sheet categories, an economically similar asset pair with different LCR weights, and the differential implementation of LCR by the very large and less-large LCR banks. We find that, since 2013, there has been reduced liquidity creation by LCR banks compared to non-LCR banks, occurring mostly through greater holdings of liquid assets and lower holdings of illiquid assets. Trends in liquid asset holdings are driven by High Quality Liquid Assets (HQLA), an LCR-defined category, particularly for assets where market and LCR liquidity weights are most similar. Of particular interest is a post-LCR shift in LCR bank portfolios to GNMA MBS rather than GSE MBS, economically similar assets with different LCR weights, that is not attributable to relatively greater issuances or relative price effects. We also find sharper declines of commercial and residential real estate loans by LCR banks relative to non-LCR banks post-2013. Finally, we find a decline in the high run-off category of LCR liabilities for LCR banks relative to non-LCR banks post-2013 for the largest LCR banks with greater than $250 billion in assets. Our results highlight the trade-off between lower liquidity creation and lower run risk from reduced liquidity mismatch of the largest banks.