Multi-criteria analysis: a manual

Multi-criteria analysis (MCA) provides a structured method for evaluating policy, program, or project options when those options cannot be fully assessed using monetary-based techniques like cost-benefit analysis (CBA). While CBA aims to value all impacts in financial terms, many public sector decisions involve social, environmental, or political impacts that are better assessed using a combination of qualitative and quantitative data. MCA bridges this gap by allowing decision-makers to weigh diverse impacts explicitly and consistently.

A typical MCA follows a systematic process: establishing the decision context, identifying relevant options, defining objectives and criteria for evaluation, and assessing the performance of each option against those criteria. By building a performance matrix—where rows represent options and columns represent criteria—decision-makers can see at a glance how various alternatives perform across multiple dimensions. More sophisticated versions, often called Multi-criteria Decision Analysis (MCDA), involve assigning numerical scores and weights to these criteria to calculate an overall ranking, allowing for a more mathematical approach to selecting the most preferred option.

The primary value of MCA lies in its ability to handle complex problems that involve multiple, often conflicting, objectives. Unlike informal or intuitive decision-making, which is susceptible to cognitive biases, an MCA provides an explicit audit trail. This transparency is crucial for public sector accountability, as it forces decision-makers to be clear about their objectives, the criteria used to measure success, and the relative importance placed on different outcomes. It does not replace human judgment but rather formalizes it, ensuring that complex trade-offs are handled in a structured, consistent, and defensible way.

While MCA is a powerful tool, it has limitations. It cannot prove that an option provides a net benefit to welfare in the way that a Pareto Improvement rule does in economics. Furthermore, the subjectivity involved in choosing criteria, scoring options, and assigning weights means the final ranking is sensitive to the inputs. Practitioners must exercise caution to avoid double-counting objectives or falling into the trap of over-complicating the analysis with too many criteria. Ultimately, the quality of the decision relies on the quality of the process design and the rigor of the underlying data.