Figuring Out Which Market Moves First Using Volatility Memory

Ever wonder if one asset's price jumps can tell you which market is leading the charge? Crypto markets bounce around constantly, making it tricky to see which asset is actually driving the action behind the scenes. I recently came across a study that looks at how researchers can rank different assets by their volatility memory. It turns out that tracking how long volatility sticks around might help uncover the hidden sequence of market moves.

The researchers wanted to figure out how to order multiple time series when an asset's own persistence masks its influence on others. To tackle this, they developed a method named QS-TCN that combines statistical volatility estimates with temporal convolutional learning. This model uses quasi-maximum likelihood and the Matrix Pencil method to set up filters that compare directional evidence between variables. That comparison creates a global order of precedence without claiming to build a full causal graph. The team tested their approach on finance benchmarks, fMRI data, synthetic systems, currency, gold and cryptocurrency markets. Across four synthetic volatility regimes, QS-TCN reached a higher mean ordering quality than an external lead-lag method, though its gain over a squared-input adaptation was uncertain in two regimes.

Curious about the details? Look up Ordering Multivariate Time Series by Volatility Memory: A Temporal Convolutional Approach to Volatility Heredity at arxiv.org/abs/2611.07746. The authors are Kaoutar El Azaar, Mohamed Hamlich and Moncef Garouani, with several coauthors.

This is not financial advice. The information provided is for educational and informational purposes only.

The cover image was generated by AI.