This master thesis analyzes the most common statistical models for measuring earningsmanagement in literature, based on a theoretical analysis and an empirical study of theGerman capital market. At first, the background and relevant motives as well as the gradualdevelopment of static models are explained, such as the model of Healy (1985) or DeAngelo(1986), towards dynamic models, the common practice in the literature. The main focus isbased on the most modern developments of dynamic models as well as those underlyingintentions. After the theoretical evaluation of the results of previous studies on the validityand reliability of those models above, they are compared with each other in an empirical studyusing a data sample of German publicly traded companies, using t-tests and multivariateregression analyzes. Moreover, this thesis examines the measurability of those modelsconcerning the effects of German enforcement mechanism by comparing the years 2000 -2005 with those of 2006 - 2012. In contrast to previous research, various models for theestimation of the extent of earnings management are used within one investigation and theirdifferent conclusions concering earnings management are analyzed and can be useful forfurther investigations. The results are very volatile and have no significant conclusions aboutthe effect of the enforcement or to a general recommendation of an "perfect" model.Individual observations of selected companies graphically show clear declines within allmodels, but they are only significant in exceptional cases. This finding is consistent with thatof Ray Ball (2013) criticisms on the reliability and validity of the models commonly used instudies on earnings management.