Quantifying Stocks Lead-Lag Relationships

The lead-lag relationship between stocks refers to the phenomenon where the movement of one stock precedes or lags behind the movement of another stock. This relationship is often analyzed in the context of stock returns and can provide valuable insights into market dynamics and investor behavior.

For instance, if Stock A consistently moves before Stock B, it could indicate that investors use information from Stock A to predict the future movement of Stock B, suggesting a lead relationship. Conversely, if Stock A tends to follow the movement of Stock B with a delay, it suggests a lag relationship.

Reference [1] quantified the lead-lag relationship by proposing a method that ranks assets from leaders to followers based on pairwise Lévy-area and cross-correlation of returns. The authors pointed out,

This paper presented a method to detect linear and nonlinear lead-lag relationships in the US equity market. In contrast to the extant literature, which uses firm characteristics such as market capitalization, trading turnover, and trading volume to select leaders and followers, our method employs the Lévy-area between pairs of stock returns to infer which one in the pair is more likely the leader, and to quantify the strength of this relationship. We constructed a portfolio that uses the previous returns of the leaders to determine positions on the followers; and showed that they generate economically significant performances that outperform all benchmarks in the literature. The performance of our portfolios is robust to various alternative specifications in algorithms, hyperparameters, and data sets…

The lead-lag relationships we find change over time. The leader-follower identity of stocks in various sectors changes several times between 1963 and 2022. This finding further supports the necessity of data-driven lead-lag detection methods that capture dynamically evolving lead-lag relationships.

Finally, we examined the performance of our portfolios across various rebalancing frequencies, and the results provided empirical support to confirm the slow information diffusion hypothesis. Specifically, the performance of portfolios decreases as both the ranking and the rebalancing are performed less frequently.

In summary,

• The paper proposed a method to quantify the lead-lag relationship between stocks.
• The relationships change, so we need to constantly use data to monitor the lead-lag dynamics.
• Long-short portfolios constructed using the identified lead and lagged stocks earn excess returns.
• Returns are better when using the daily timeframe. This validates the information diffusion hypothesis.

Let us know what you think in the comments below or in the discussion forum.

References

[1] Cartea, Álvaro and Cucuringu, Mihai and Jin, Qi, Detecting Lead-Lag Relationships in Stock Returns and Portfolio Strategies (2023). https://ift.tt/iyBPdQY

Originally Published Here: Quantifying Stocks Lead-Lag Relationships

source https://harbourfronts.com/quantifying-stocks-lead-lag-relationships/