Modifiable Areal Unit Problem (MAUP) is a widely known issue in fields like epidemiology, political science, and economic geography. In this session we will introduce and discuss this topic in the context of insurance, specifically territorial ratemaking and monitoring.

For many geographic studies, we need to aggregate spatial data from the smallest units (street-level address, lat-long) to some territorial groups (preferably contiguous, e.g., zip codes, census blocks, etc.). This required spatial grouping leads to the problem known as MAUP. As described by S. Openshaw in his 1984 paper, the origin of MAUP is “the areal units (zonal objects) used in many geographical studies are arbitrary, modifiable, and subject to the whims and fancies of whoever is doing, or did, the aggregating.” In insurance context we see this phenomenon often times as there is no standard procedure for choosing territorial boundaries. Most recent trends suggest using zip codes or census blocks; both of which are actually devised for non-insurance purpose and may not be optimal for insurance ratemaking purpose. More importantly, territorial base rates may change drastically with the choice of territorial boundary definition.

There are two main elements of MAUP: First, scale problem—which deals with the variation caused by the number of territories used in the analysis. Cost level of a representative territory should be close to its member insured. Larger territories can fail to preserve the property of its members. This is also known as ecological fallacy. Recent developments in insurance have addressed this problem to a great extent by using granular territorial definitions (zip codes, census block, etc.). However, the second element of MAUP, aggregation problem, is very much present and needs to be addressed in insurance ratemaking. Aggregation problem arises when there exists significant variations in results due to the use of alternative territorial boundary definitions (with same number of units). Different territorial definitions generate different territorial base rates for the same insured, which may be even different from insured’s true base rate.

The speakers will use most of the session discussing and understanding the relevance of MAUP in insurance ratemaking. They will spend the rest of the time proposing a different techniques that will help making better statistical inference and business decision in the existence of MAUP.