Relating Portfolio Turnover to Risk and Signal Persistence

Summary

The document raises the problem of interpreting portfolio turnover when market stress causes trading activity to rise. It asks whether a turnover threshold should be adjusted for market or portfolio volatility, perhaps using a benchmark and beta, while recognizing that high turnover during stress may itself be informative. It also notes alternative activity and holding-period measures, but argues that they do not resolve the issue for the author.

As a possible starting point, the question cites a published expression relating turnover to the number of eligible stocks, target tracking error, average inverse stock-specific risk, and the first-order persistence of the alpha signal. The author says this formula does not directly provide a practical threshold and asks how it might be adapted. No answer, empirical test, or recommended normalization is included, so the document frames a research problem rather than establishing a method. Any threshold would need to clarify whether it is intended to control trading costs, characterize strategy behavior, or evaluate performance across regimes.

Key ideas

• Market stress can coincide with sharp increases in portfolio turnover and complicate interpretation.
• The question considers adjusting turnover for volatility, potentially using a benchmark and beta.
• A cited model links expected turnover to tracking error, stock-specific risk, universe size, and alpha signal persistence.
• The document proposes no tested threshold or solution, leaving the adjustment as an open research question.

Tags

# How to set a portfolio turnover ratio threshold according to volatility?


# How to set a portfolio turnover ratio threshold according to volatility?












With the various crises affecting the financial markets, here a pandemic, this ratio skyrockets almost every time, which limits its interpretability.

We know that the ratio increased during the subprime financial crisis, for example, and that it continuously decreased thereafter, until we were affected by Covid-19, Brexit and US elections. So with all this things happening it should be quite normal that managers would make up their mind and just shift in the strategy of the funds.

I'm aware of modified turnover rate, churn rate and Cremers and Pareek duration measure, but I don't think they solve it.

So is that a solution would be to adjust the ratio to the volatility (index or whatsoever) ? Or would it lose its purpose (because after all a study prove that it's most relevant in times of stress) ? I thought about using a benchmark and beta.

Qian, Hua and H. Sorensen (2007) gave a bit of a hint with the following formula :

$TO = \sqrt{\frac{N}{\pi}}\sigma_p \sqrt{1-\rho_z(1)} E_{cs}(\frac{1}{\sigma_i})$,

where : $N$ number of stocks in selection universe, $\sigma_p $ : portfolio’s target tracking error, $\sigma_i$ : stock i’s specific risk, $\rho_z(1)$: first order alpha signal correlation.

It's obviously not really appropriate for setting a threshold but there is an idea, I don't really know how to adapt it, someone could help please ?

Thanks !

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.