Accounting Anomalies And Stock Return Prediction

We apply Penman’s model on Chinese data to predict stock returns withaccounting anomalies. Considering differences between US and China in accountingstandards and practice, we modify the variables and their components; for example,we exclude other receivables and other payables, as often accused of hiding relatedparty fund occupation. Our sample covers2001-2012A-share listed Chinese firms,excluding financial firms. The regression shows similar results as in US, withsignificant abnormal return in out-of-sample test on a five year rolling basis. Forcomparison, we also include other receivables and payables to verify the result.Furthermore, we extend our research to several different conditions. Firstly, weshorten return calculation period to April30thtill August31st, instead of yearlyinterval, and find release of interim report has little effect on investors’ decision.Besides, we classify companies into industry groups and regress on each group, onlyto find obvious divergence between energy industry and others. Thirdly, as lossmaking firms are often liable to accounting manipulation and might reduce anomalyeffect, we test profitable firms sample and find no improvement in abnormal return.Thus profitable firms do not show stronger accrual anomaly.