Banks use trading as a vehicle to take risk. Using high-frequency regulatory data, we estimate the sensitivity of weekly bank trading profits to aggregate equity, fixed-income, credit, currency, and commodity risk factors. Our estimates imply that U.S. banks had large trading exposures to equity market risk before the Volcker Rule, which they curtailed afterwards. Credit and currency risk exposures were smaller. The results hold up in a quasi-natural experiment that exploits the phased-in introduction of reporting requirements for identification. Total trading-book risk was also curtailed afterwards with material financial stability implications and no evidence of migration to other bank activities.

We study dealers’ bidding behavior in the Bank of England’s quantitative easing (QE) reverse auctions. Using a granular dataset on both accepted and rejected offers together with an equilibrium model of bidding behavior, we estimate dealers’ valuations of securities offered to the Bank of England. We also recover the rents accruing to dealers from participating in the auctions as opposed to liquidating gilts in the secondary market, thereby possibly causing prices to change. These rents or so-called ”liquidity benefits” are largest in the early phases of QE implemented during the Global Financial Crisis, suggesting that QE may be particularly effective in restoring smooth market functioning when market participants are facing large liquidity shocks. Finally, we document that dealers’ valuations vary significantly with the amount of interest rate risk acquired in the secondary gilt market before the auction and with dealers’ regulatory capital.