Working Paper: CEPR ID: DP9958
Authors: Stefano Rossi; Katrin Tinn
Abstract: We present a model of quantitative trading as an automated system under human supervision. Contrary to previous literature we show that price-contingent trading is the profitable equilibrium strategy of large rational agents in efficient markets. The key ingredient is uncertainty about whether a large trader is informed about fundamentals. Even when uninformed, he still learns more from prices than market participants who still wonder about whether he is informed. Therefore, he will trade a non-zero quantity based on past prices, whose direction ? trend-following or contrarian ? depends on parameters. When informed, he will trade on that information and disregard the algorithm. One implication is that future order flow is predictable even if markets are semi-strong efficient by construction.
Keywords: Kyle model; Log-concavity; Rational expectations; Rational price contingent trading
JEL Codes: D82; G12; G14
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| type of trader (informed or uninformed) (G14) | trading behavior (G41) |
| trading behavior (G41) | market prices (P22) |
| uncertainty about large trader's information (D89) | information advantage (D83) |
| trading behavior (G41) | predictable future order flows (C69) |
| probability of no news (D80) | direction of trading (trend-following vs. contrarian) (G41) |
| size of order flows (C69) | direction of trading (trend-following vs. contrarian) (G41) |
| trading strategies (G13) | market efficiency (G14) |