The present study aimed at investigating the effect of Ethical Leadership (EL) on Employee Performance (EP) at private and governmental commercial banks in Bahir Dar city branches. The main objective of the study was to show the effect of EL to EP. Crosssection descriptive survey research strategy was adopted in which 45 usable structured close-ended questionnaires were collected from 152 distributed respondents. The independent variable ethical leadership was measured by ELQ consisting of 15-items rated on 5-point Likert scale adopted from Yukl et al. (2013). The dependent variable employee performance was measured by 6-items developed by Tsui et al. (1997) rated on 5-point Likert scale. The mediating factors Organizational Citizenship Behavior and Counterproductive Work Behavior were measure by 10-items each rated on 5-ponit Likert scale as adapted from Spector et al. (2010). The descriptive and inferential statistical techniques were used for data analysis. In the inferential statistics, Pearson’s correlation and regression analysis were used to assess both relationships and effects for the hypothesis formulated in the study. The findings have shown that; first, there exist positively significant relationship between ethical leadership and employee performance indicating that, the employees with high perception of ethical leadership are likely to exhibit better performance. Second, organization citizenship behavior doesn’t significantly mediate the relationship between ethical leadership and employee performance. Third, counterproductive work behavior doesn’t significantly mediate the relationship between ethical leadership and employee performance. Results of other hypotheses were presented in chapter four. From seven hypotheses, five of them were supported by existing research findings whereas two of them are novel ones. All findings were considered in the context of perceptions of employee of private and governmental commercial banks in Bahir Dar City branches leading the researcher to consider and test under other cultures, situations and environments for novel results inconsistent with existing literatures. Analyses were carried out using IBM SPSS Statistics version 20 and IBM SPSS AMOS version 21.