Fitting Climate Catastrophe Aggregate Loss Distributions: A Case Study of the Fictional Country of Storslysia
Open Access
Author:
Magulick, Lacey
Area of Honors:
Actuarial Science
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Zhongyi Yuan, Thesis Supervisor John Lesieutre, Faculty Reader Amanda W Hammell, Thesis Honors Advisor
Keywords:
actuarial science statistical modeling aggregate loss climate
Abstract:
This paper is an extension of the 2023 Society of Actuaries (SOA) Student Research Case Study Challenge that I participated in with fellow Schreyer scholars. I begin by giving an overview of the case study and providing an outline of our final submission. The case study is centered around a fictional country, Storslysia, and crafting a social insurance program for the nation to prepare for future natural disasters. I aim to redo the model that we created for the case study to make it more advanced and achieve a better estimate of aggregate loss. I then move into my literature review which includes analyzing the top placing submissions from the case study, along with research on different actuarial and statistical models that were considered in the revised modeling process. Next, I break down and explain my modeling approach beginning with categorizing every historical weather event into a hazard type category and conducting exploratory data analysis. I then analyzed frequency and severity of the different hazard types and individually fit a statistical model to each. I explain what models were considered and ultimately selected as the best fit for both frequency and severity with justification. I list my data assumptions and limitations and how I achieved the aggregate loss for the country of Storslysia using the new model. Lastly, I conclude the paper by comparing my advanced actuarial model to the simpler model presented in the case study submission.