Knowledge that Transforms

Prediction markets are a successful information aggregation structure, however the exact mechanism by which private information is incorporated into the price remains poorly understood. We introduce a novel method based on the “Kyle model” to identify traders who contribute valuable information to the market price. Applied to a large field prediction market dataset, we identify traders whose trades have positive informational price impact. In contrast to others, these traders realize profit (on average) in excess of a theoretical expected informed lower bound. Results are replicated on other field prediction market datasets, providing strong evidence in favor of the Kyle model.

Does media coverage of a firm have a causal effect on the volatility of its stock price and, if so, is this of aggregate importance? I identify a robust link between coverage in the Financial Times and a firm’s intraday stock price volatility. This effect is not driven by persistence in volatility or anticipation of future newsworthy events, but is explained by an increase in trading volume, supporting a salience interpretation. The effect spills over into firms related by the structure of the production network, but does not affect the aggregate level of volatility.

We test the joint efficiency of the bitcoin and ether options and perpetual futures markets and identify the determinants of arbitrage opportunities. Our novel fiat-currency-free put–call parity relationship motivates new arbitrage tests for options-only and option–perpetual cross-markets. Bitcoin and ether derivatives markets are becoming more efficient, especially for options of maturity ≥ 15 days. Bitcoin derivative markets are generally more efficient than ether derivative markets, but arbitrage strategies can still be highly profitable even under conservative transaction cost scenarios, which include slippage for large orders, especially during periods of high trading volumes or when the blockchain traffic becomes more congested.

We study the impact of social media posts and news articles on retail trading on Robinhood by analyzing the net buying activity before and after a (social) media publication. Our findings indicate that both publication types lead to an increase of holders. However, both effects are short-lived, and the effect of social media posts on retail trading is significantly higher. Furthermore, investors buy stocks that are recommended on Reddit regardless of text sentiment. However, they act as contrarians by selling stocks with positive news coverage and vice versa. We thus find evidence for both attention- and information-driven trading behavior.

I study the effect of algorithmic trading on market efficiency, taking into account past market and limit order flows alike. I find that an exogenous increase in algorithmic trading around the introduction of the NYSE Hybrid Market leads to a significant decrease in the predictive power of surprises in market order imbalance and limit order book imbalances, especially at the outer levels of the limit order book. However, the predictive power of past returns remains largely unchanged. This suggests that algorithmic trading improves market efficiency by facilitating the incorporation of information embedded in both market and limit order flows.

We analyze the lead–lag relation between VIX futures and SPX futures. The two futures markets are weakly connected when market volatility is low. By contrast, when volatility is high, their prices are highly negatively correlated, with VIX futures leading SPX futures. However, the tightness of the lead–lag relation prevents the formation of profitable trading strategies in a setup that includes transaction costs. An analysis of the time variation in the lead–lag relation finds that an improvement in the relative liquidity of one market strengthens the lead of that market. Moreover, the hedging activities of market makers influence the lead–lag relation.

I demonstrate empirically that corporate bond dealers mitigate adverse selection risk by passing potentially informed transactions to institutional investors. I contrast price reversals following days with abnormal trading volume across bonds with different information asymmetry. In informed trading, the part of reversal specific to high-volume days should increase with information asymmetry. In uninformed trading, there is no such effect. Following high-volume days when investors provide liquidity, the reversals are consistent with the former case. When dealers provide liquidity, I observe the latter. The results suggest that the informational content of bond prices is higher when dealers do not take inventory.