Traders tend to opt for long trade durations to minimize aggregate transaction costs over many trades. In this research I show that this is not necessarily optimal when short term stock price information is available. I show that price movement during a trade explains up to 60% of the variation of transaction costs. Furthermore, I show that we can use short term information to capitalize on this. By changing the duration or timing of trades according to machine learning predictions using logistic regression, neural nets, and LSTM’s we can improve transaction costs up to 32 basis points per trade. This amounts to a net gain per transaction which generally does not consistently occur.

Additional Metadata
Thesis Advisor Vermeulen, S.H.L.C.G.
Persistent URL
Series Econometrie
Supit, S.J. (2020, July 14). Transaction Costs and Short Term Price Signals: a Happy Marriage. Econometrie. Retrieved from