5 Comments
User's avatar
Thomas's avatar

Interesting concept. Potential improvement would be including delisted stocks, and point in time macroeconomic data.

In addition, transaction cost could impact time series momentum that go long for all stocks with positive momentum could include many stocks. Perhaps using a cutoff (eg only long top N momentum stocks) can be a practical alternative?

Expand full comment
Nam Nguyen Ph.D.'s avatar

Going long all the stock would not be practical for smaller portfolios. Implementing a turnover control mechanism can make this strat more practical

https://harbourfrontquant.substack.com/p/does-trend-following-still-work-on

Expand full comment
Thomas's avatar

Thanks for the pointer Nam. Interesting paper 🙏🏻

Expand full comment
Neerav Mehta's avatar

Results are on back-tested period. There is no test or validation dataset.

Expand full comment
Nam Nguyen Ph.D.'s avatar

yes, only IS data

Expand full comment