Corporate earnings announcements unpack large bundles of public information that should, in efficient markets, trigger jumps in stock prices. Testing this implication is difficult in practice, as it requires noisy high-frequency data from after-hours markets, where most earnings announcements are released. Using a unique dataset and a new microstructure noise-robust jump test, we show that earnings announcements almost always induce jumps in the stock price of announcing firms. They also significantly raise the probability of price co-jumps in non-announcing firms and the market. We find that returns from a post-announcement trading strategy are consistent with efficient price formation after 2016.

We study the trading activity of designated market makers (DMMs) in electronic markets using a unique dataset with audit-trail information on trader classification. DMMs may either adhere to their market-making agreements and offer immediacy during periods of heavy selling pressure, or they might lean-with-the-wind to profit from private information. We test these competing theories during extreme (downward) price movements, which we detect using a novel methodology. We show that DMMs provide liquidity when the selling pressure is concentrated on a single stock, but consume liquidity (leaving liquidity provision to slower traders) when several stocks are affected.