Memory and Trading

Abstract

I test the predictions of human memory models in a high-stakes trading environment. Using alphabetical rankings of stocks from portfolio statements, I estimate plausibly random associations of adjacent stocks in an investor’s memory. When two stocks are associated in an investor’s memory, trading one stock cues the recall of the other, and increases the probability that the investor also trades the other stock. Increasing the memory strength of this association by one standard deviation increases the trade probability by about 5 percentage points. I then document that personal experience affects trading behavior through the different properties of human memory. My results help uncover the sources of experience effects and provide guidance for models of memory and financial decision-making. My results also demonstrate how theory-guided tests can uncover new facts about investor behavior.

Keywords: Memory, Trading, Household Finance, Mutual Funds

JEL Classification: G41, G11