Multilevel modeling is a powerful tool for dealing with complex data structures including hierarchical and cross-classified datasets.  Thus, it can easily accommodate for many of the special features of insurance data (especially group insurance).  It can also allow for panel/longitudinal aspects of the data, which are often important especially for experience rating purposes.  

Despite the significant importance of this modeling approach and its high potentials for actuarial applications, it started to find its way to actuarial literature only recently.  This may be partly attributed to the unfamiliarity of this technique to many actuarial researchers and practitioners.  Therefore, the key purpose of this paper is to provide a clear & concise introduction to this important topic within an actuarial context. 

In particular, this paper is written as an introductory primer suitable for both academics & practitioners.  It starts from the familiar (single-level) general linear model.  Then, it builds up to the more advanced multilevel modeling approach, with all new terms clearly explained.  The paper also highlights the key benefits gained by using this approach, and relevant recent applications from insurance and actuarial literature.  Interested readers are also referred to more elaborated references for further reading.  Overall, it is hoped that this paper will encourage more researchers and practitioners to adopt this important technique in their future work.

*Awarded 2nd Place Non-Life Track Prize