Microsoft Word 2007 ichrie conference Proceedings Final-Final 06-06-07. doc

 
Figure 1
EM strategies illustration. (revised from Mohanram (2003))
 
Seminal research in this area is from Beneish (1997) who worked in the development of earnings detection 
models that is the focus of our present research for the lodging industry. Beneish (1997) develops a model that 
distinguishes actual GAAP violators sanctioned by the SEC from merely “aggressive accruers”. However, the 
model is really only applicable to those firms experiencing extreme financial performance, presumably poor. 
Beneish (1999a) further adapts his earlier model in an attempt to devise a new model that would detect if a 
firm was manipulating earnings, not just those experiencing extreme financial performance as in his 1997 paper.
The model uses a dichotomous dependent variable and a matrix of eight independent explanatory variables. A 
description of the variables and their calculation is shown in the appendix.
After compiling the matrix of variables, Beneish (1999a) first compares the eight separate variables using 
both parametric and nonparametric methods. He finds that firms that engage in earnings manipulation differ 
significantly from those that do not in terms of receivables, gross margins and asset quality. He also finds that 
manipulating firms have higher sales growth and have a higher ratio of accruals to total assets. 
His research also includes the use of weighted exogenous sample maximum likelihood (WESML) and 
unweighted probit models to determine which variables significantly influence earnings manipulation. While both 
models have similar predictive validity, he focuses on the results of the unweighted probit model. He finds the 
significant variables to be days’ receivables, gross margin, asset quality, sales growth and total accruals to total 
Target 
Big bath 
Bump up 
Cookie jar 
Earnings smoothing 

---