The Principal-Agent (PA) Theory outlines the relationship between two actors in an implementation process, wherein the principal (in this case: policy makers) makes policy decisions they want the agent (street-level bureaucrats) to execute. The Principal-Agent theory understands implementation in a technical and idealistic ‘top-down’ approach, but misses a few key considerations for understanding the implementation process. It assumes that there is always an information asymmetry between the principal and the agent, which results in policy implementation conflict in terms of the policy outcomes and policy impacts. Bottom-up policy implementers, on the other hand, associate the agents (SLBs) to be just as good as the policy they implement, as the policy’s target audience would associate the success of the policy to the way the agents make decisions, establish routines, and ultimately process implementation. The PA Theory sees the agents as tools to deliver the goals of the principal, and when the policy does not achieve its intended aims, it is assumed that there is an erroneous practice in the part of the agent. The Principal Agent theory in public policy is often used to rationalize gaps in unsuccessful (or the lacking of success of) implementation process in favour of the top-down approach. The theory assumes that a rational act entails the execution of the principal’s decisions through agent’s action. However, it does not take into account other factors, which will be further discussed in the essay, that may lead to the agents taking differing courses of action, and conflict-ridden divergences of policy implementation. The PA model negatively perceives divergence, assuming that it is solely the result of an agent’s self-interested motives, and hence a ‘failure’ in the implementation of the policy. In this essay, I will argue that the PA theory is inadequate in modelling and understanding the policy implementation conflict from a bottom-up implementation approach. I will mainly discuss the ways in which SLBs’ discretion, emotion, situational and contextual reasoning fosters greater policy legitimacy, democratic legitimacy, and policy change that would be most fitting to the society in which the policy is implemented, and is not always as technical (and linear) as the PA theory assumes. However, although I argue that the PA theory has significant drawbacks in looking at the principal-agent relationship, I do not disregard the notion that the PAT has important and relevant points to make in the debate, despite its issue with one-dimensionality and parsimony that the>