Retesting Penman, Richardson and Tuna’s theories

Introduction

In this paper, we will retest Penman, Richardson and Tuna’s theories and hypothesis using UK evidence, and try to find whether the data from UK listed company can support their theories. Some arguments from Piotroski will be discussed in this paper, while the accounting system is also considered as a crucial factor which may affect our testing result.

Methodology

Corrleations

Basic Correlations

The correlation table is a basic method to show whether the B/M effect is significant, and the relationships between different variables and future returns. To observe how the variables in PRT’s model are in associate with future returns, we divide the firm-year observations into 10 portfolios in each correlation tables by firm-year book-to-market ratio, operating component of B/M ratio ( , financial leverage component of B/M ratio ( , and B/M minus operating component ratio ( .

The returns are calculated as buy and holding stock year return. It covers the next 12 months return which begins the first month of the next fiscal year. For most of the companies publish the annual reports in March or April, there is a time gap between the future return and market reaction. In our opinion, the first four months stock returns are affected by the expectations of the companies’ performance, thus we use the next whole year as the future stock returns in this paper. Because stock return for year 2010 is not available when we obtain the data; we exclude the data for year 2009 in correlation analysis. Since the dividend occupies only a small part of the returns, the dividend return is ignored in our correlation analysis to simplify our calculation. The negative data for variables for B/M ratio, operating component and ...

... middle of paper ...

...relation analysis in Table 2. When NOA/P ≧1, leverage is positively relative to the NOA/P; however, when NOA/P﹤1, leverage is negatively relative to NOA/P.

Regression V shows the leverage coefficient under controlling of operating risk (enterprise B/M ratio). For full sample testing, the coefficient is insignificant which means we cannot get any reliable conclusions from this result. Nevertheless, for NOA/P ≧1, the coefficient is significant positive, and for NOA/P﹤1, the coefficient is significant negative. If we split the ND/P ratio into financial liabilities/P and financial assets/p like in regression VII, FA/P coefficients are significantly positive in all of three panels, but the FL/P negatively or insignificantly relative to future returns. It indicates that the high future return premium is awarded for the high operating risk rather than financial risk.