Using Mathematics to Quantify Subjective Decisions: Application of Analytic Hierarchy Process to Risk Assessment

Abstract

Risk management requires human judgements, from risk identification, assessment to response. Although automated tools are useful in handling large amounts of data and in performing complex calculations rapidly, humans undertake the entire risk management process. They bring to the process their intuitions, insights, previous experiences and skills. Therefore, creating a rich source of information of risks faced by an organisation. Ignoring human factors may impoverish information and limit risk management to only measurable factors. This study contributes to the field of decision-making and risk assessment by investigating and discussing in detail how to quantify subjective judgements using the Analytic Hierarchy Process (AHP). AHP is used to assess risk of an insurance company. It discusses how to do risk assessment by combining both intuition and analytic in the decision-making process. The study defines intuition as knowledge and experience, and analytic as the mathematics or quantitative analysis to derive the result. It demonstrates how Analytic Hierarchy
Process (AHP) - a flexible multi-attribute or multi-criteria decision making tool, enables risk managers to use both intuition and analytic to do risk assessment. Risk assessment using AHP produces global priority weights representing the overall risk ranking of an insurance company. The study develops a risk assessment problem and uses AHP to organise and structure risks and sub-risks of the problem. It uses formative evaluation method with open-ended questionnaires to obtain feedbacks from risk managers on AHP. Three employees of a risk management department in a government agency assesses the risks using AHP. AHP strengths are easy to use and understand, improves risk assessment and useful for risk assessment problems that have scarce or no data. AHP limitation are the numbers and repetitiveness of the pairwise comparisons. The participants either ignore some of the pairwise questions or they answer randomly instead of deliberate judgements.