The Existence And Robustness Of Accrual Anomaly

Inspection on efficiency of capital market efficiency has been a hot topic in the field of financial research in the past few decades. As the basis of modern financial research, efficient market theory has important theoretical and practical value. However, since the1980s, many scholars have found a lot of anomaly which cannot be explained by efficient market and capital asset pricing model. These emerging anomalies launched a strong impact on effectiveness of market, and attracted lot of researchers.The accrual anomaly discussed in this thesis is a kind of market anomaly. It is discovered and named by Sloan in1996. Sloan studied reflection of the stock price on the structure of earnings and then proposed accounting accruals negatively correlated with stock returns, which confirmed the arbitrage strategy can get an average10.4%return. However, as a hot topic, accrual anomaly has been studied by lots of researchers home and abroad in the past few years, they failed to come to a consistent conclusion. By reviewing the previous literature, we could find although in different studies researchers had basically comply with Sloan’s research ideas, they used slightly different research design. Therefore, this thesis attempted to compare impact of different research design on the research conclusion, analyzed relationship between existences of accrual anomaly and differences of research design. Then it explained the reasons for why different conclusions were made in researches on accrual anomaly existence and illustrated the importance of selecting an appropriate research method.Specifically, this thesis focused on four aspects of research design, including the definition and measurement of accrual, the adjustment to the returns, the impact of the different method dealing with outliers, as well as deciding weights while constructing the portfolio. In every aspect, we chose widely used and representative definition and calculation method in most studies. And in empirical research, test results using different methods of measurement and processing were reported. According to the analysis of the empirical results of different research methods, we get the following main conclusions:Firstly, earnings persist as a whole. When we divide earnings into cash flow and accrual, persistent of cash flow is significantly higher than that of accrual. Secondly, the different approach dealing with outliers is the most critical factor affecting the ultimate returns of portfolio. Thirdly, by comparing with prior researches, we believe differences in research design have led to contradictory conclusions of different researches. Finally, we believe that the existence of accrual anomaly relies heavily on research methods. In most cases, we cannot directly observe key evidence of existence of accrual anomaly. Thus we speculate that the market may have well reflected the differences of persistence between cash flow and accruals.The possible contributions of this thesis are as follows. Firstly, we focus on the importance of research design. We illustrated the influences of different research design on the final result with comparison, and then reminded the researchers of the importance of selection of research methods. Secondly, we explained the reasons for contradictory conclusions of prior related researches from the point of view of differences in research design. Thirdly, we used data generated recently, and the time span covered is also longer, thus this thesis overcome the defects of previous study:under the new guidelines, data is not enough and time span is short.Possible defects and deficiencies are as follows. Firstly, we only selected some widely used and relatively representative method of measurement and data processing, while we didn’t discuss research designs which were used in fewer cases. Secondly, we didn’t conclude which research design is better. In addition, we only considered returns of portfolio within one year after the publication of the financial statements, thus didn’t take long-term performance of the portfolio into account.