A risk management professional at a bank is attempting to determine the distribution of potential losses stemming from operational risks (e.g., internal fraud, system failures, and other disruptions). However, the professional is facing a challenge due to the limited availability of historical data on these types of losses. Considering this constraint, what is the most effective approach to accurately estimate the severity of these operational losses?