Commonality in Two-Dimensions: An Empirical Investigation

<p>In this thesis, I follow Hasbrouck and Seppi (2001)’s work and use reduced-rank regression to model the commonality in Chapter Two. The literature on the study of return commonality generally attributes its source to the order flow. But I find that return and order flows are endogenous and use the new exogenous Twitter sentiment dataset to show that return commonality may be due to sentiment and attention. Furthermore, I observe the non-linear (linear) relationship between sentiment (attention) and return commonality. Finally, I may export the non-linear relationship using the same reduced-rank regression framework in future research.</p> <p>I also follow Korajczyk and Sadka (2008)’s work in Chapter Three. They use PCA to extract a systematic liquidity factor from eight liquidity measures and show that it is a priced factor. Previous studies conceptually state that liquidity has multiple dimensions (three to five dimensions). Still, few studies discuss how to extract one or a set of systematic liquidity factors from various liquidity measures. I fill this gap and empirically confirm that the systematic liquidity factor is multi-dimensional and priced multi-dimensionally using the multi-linear PCA (MPCA) method on daily-level data. MPCA allows me to provide a nice explanation of the factor loadings of the systematic liquidity factor. In future work, I want to apply the MPCA method to asset pricing with large dimensions of firm characteristics.</p>

Name thesis:degree_name

Doctor of Philosophy (PhD)

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Finance

Year

2022

Author dc:creator

• Zhou, Zhaoque

Contributors dc:contributor

• Raja Velu
• Lai Xu

• Commonality
• High-frequency
• Liquidity
• Multi-dimension
• Business
• Finance and Financial Management

Repository record dc:identifier

https://surface.syr.edu/etd/1623

OAI identifier oai:identifier

oai:surface.syr.edu:etd-2624