This study examines the washback effect of a final examination. Despite the general criticisms of a final examination for its negative influence on education, no empirical evidence was noted in the existing literature; rather some evidence for positive washback was found. The study was based on the assumption that the detrimental effect of a final exam is not inherent; whether the washback is negative or positive is dependent on the design of the exam. Furthermore, the power that the exam has to influence teaching and learning, if exploited properly, can make it work as an instrument for educational change. The context of the study was English language teaching and testing at school in Nepal. The washback effect of the School Leaving Certificate English exam, an exit exam based on prescribed textbooks in terms of its content and discrete-point approach in terms of its format, was examined. A new exam of reading, grammar, note-taking and writing based on the course objectives of the SLC English course in terms of its content, and integrative-communicative approach in terms of its format was designed to use as the criterion measure of the English proficiency of the students. The new exam was administered to school leavers and students of the previous year. The performance of the school leavers on the new exam was compared with their performance on the SLC English exam, and with the performance of the students of the previous year on the new exam. The results indicate that SLC English exam had a negative washback on the teaching and learning of the SLC English course because it failed to allow the students and the teacher to work for the course objectives of the SLC English. It was concluded that washback is an inherent quality of a final exam; people whose future is affected by the exam-results work for the exam regardless of the quality of the exam. Whether the washback is negative or positive is dependent on what the exam measures; if it is congruent with the sentiment and the purposes of the course objectives, it can achieve beneficial washback; if not it is bound to produce harmful washback. Innovations through the former type of exam would lead the teaching for the exam to be in accordance with it. The implications for language testing in general, and the Nepalese ELT situation in particular are presented. It is suggested that the SLC English exam should be replaced by an exam similar to the one used, in order to bring about change in the teaching of the SLC English course. Recommendations for further research are made.