Value-added models try to separate the contribution of individual teachers or schools to students' learning growth measured by standardized test scores. There is a policy trend to use value-added modeling to evaluate teachers because of its face validity and superficial objectiveness. This article investigates the potential long term consequences of making high-stakes decisions based on value-added teacher evaluations. To investigate this question, we analyze the micro-level effects on teacher effectiveness from the view of policy implementation and the macro-level effects on teacher quality based on the dynamic change of the teacher job market. We argue that the establishment of a formal connection between value-added teacher evaluations and high-stakes decision-making may compromise teacher effectiveness and teacher quality. We conclude that this connection between value-added measures and high-stakes decisions should not be established when it compromises the perception of a teacher's position as a secure and decent job.