multifactor valuedog screen
The top 50 firms from this screening algorithm are reported. The only requirement to be included in the initial population is that the market value exceed $100 million. Each variable is defined so that a higher value is better. The precise definitions of the five screening variables are given below:
 |
Surprise = (actual quarterly EPS - forecast quarterly EPS)/price at the end of the fiscal quarter, where both the actual and forecast amounts are from FirstCall. |
|
ProForm Exclusions = (GAAP EPS - actual FirstCall EPS)/total assets at the end of the fiscal quarter. GAAP EPS is from Compustat and is diluted and before Extraordinary Items. |
|
Financial Health Score is between [0,9] and is described on here. |
|
Price-to-Book = Price / book value at fiscal quarter end. Because the Financial Health Screen only applies to firms with the highest 20% price-to-book ratio, I also screen on this ratio to bring more of these firms to the top of the sort. |
|
Negative Accruals = - (Net Income before Extraordinary Items - Cash from Operations)/Total Assets at the end of the fiscal quarter. |
|
Residual Income Valuation = [book value per share + 1.3*(FirstCall next 12 months EPS estimate - .10(book value per share))]/price |
|
Momentum = Six month compounded stock return. While not itself a financial statement "fundamental," this variable appears to help isolate the winners in the top Surprise decile (but this conclusion is based on as yet unpublished research). |
The multifactor screen sorts the population by each variable and gives each a decile rank (with 1 being the top 10% and 10 being the bottom 10%). If the data is missing it is assigned a rank of 5, with the exception of the Financial Health Score which only applies to a subset of the population. The reported average score on the far right of the spreadsheet is the average of the rankings, so a smaller value is better.
See research papers for the science behind each of these variables.