Improving Trend Estimation Using Mix of Business Data

# MotivationWhen using both premium and loss trend data for ratemaking, smoothing techniques are often used to remove random noise to discover the "true" trends. By using the detailed granular data available for many lines of business, a more comprehensive analysis is possible, which allows for a more nuanced result. In addition to helping explain what is driving trend changes, the detailed data can also provide an early warning of potential adverse selection and valuable insights on how well a rating plan is performing.# MethodThis paper will examine how changes in the mix of business can impact the observed trend data. In addition, ways to account for the mix of business changes are suggested in order to better understand the underlying trends.# ResultsChanges in the mix of business over time can account for a significant portion of the trends in observed data. By adjusting trend data for changes in the mix of business, the residual or "true" trends are easier to identify.# ConclusionsNot all ratemaking data can be adjusted for changes in the mix of business. When the data can be adjusted, there may be a significant improvement in both the transparency and accuracy of rate indications. In addition to improved trend estimation, this process can also lead to a better understanding of a book of business and provide an early warning system for potential adverse selection.# AvailabilityAll data and examples used in this paper are available for review in the companion Excel file. While useful to explore concepts discussed in this paper, it is not necessary to review the Excel file before reading the paper.