Cluster Analysis on PEAD Using SUE Model with Quarterly Data

Huaiyu Sun ,Mi Zhu ,Feng He

doi:10.4236/ib.2013.51b007

Abstract

The paper constructs SUE model with Quarterly Report Data of the A-share listed companies for the first time, to analyze Post-Earnings-Announcement Drift (PEAD) with a higher data frequency. A further cluster analysis is made to examine PEAD phenomenon under different market caps and report periods. The study finds that large cap stocks show stronger PEAD than the small cap. It also firstly finds that PEAD on Q1 is more obvious than others. These findings have practical significance for the investment management decisions and the quantitative trading financial products developments.

Highlights

Ball and Brown (1968) firstly found “Post-EarningsAnnouncement-Drift”; that is, stock prices show continuous and enduring reaction to the latest information, which means stable excess return can be obtained through quantitative investment model based on excavating on report data. Bernard and Thomas (1989) demonstrated that the Post-Earnings-Announcement Drift (PEAD) is attributed to the CAPM model’s mispricing to the market. Bhushan (1994) considered PEAD, directly or indirectly, driven by the transaction cost
Decile 10 represents the largest group of Standardized unexpected earnings (SUE), which means the highest unexpected earnings; Decile 1 represents the smallest group of SUE. 0 on abscissa represents the announcement date
The SUE model with quarterly report data validates PEAD driven by unexpected earnings

Summary

Introduction

Ball and Brown (1968) firstly found “Post-EarningsAnnouncement-Drift” (shorts for PEAD); that is, stock prices show continuous and enduring reaction to the latest information, which means stable excess return can be obtained through quantitative investment model based on excavating on report data. Bernard and Thomas (1989) demonstrated that the PEAD is attributed to the CAPM model’s mispricing to the market. Bhushan (1994) considered PEAD, directly or indirectly, driven by the transaction cost. Standardized unexpected earnings (SUE) model was firstly presented by Rendleman, Hones and Latane (1982), and the model was used to measure PEAD by Bernard and Thomas (1989), who pointed out that the drifting magnitude is more obvious in small-cap firms than the large ones. This paper constructs SUE model with quarterly report data of A-share market for the first time, and use cluster analysis to estimate PEAD of different market caps and periods. The study makes up for the low sampling frequency deficiency of the previous research, and finds out two new features of PEAD in A-share market. These findings help to deepen the understanding of the nature of A-share market, and provide the basis and reference for the development of quantitative financial innovative products.

Model Design

General Experiment Results

Cluster Analysis and Results of Different Market Cap

Cluster Analysis and Results of Different Report Periods

Conclusion