An Econometric Analysis of Paid Sewer Backup Damage Claims In Seattle.

摘要

Seattle is known for its high occurrence of rainfall events but most of them are low intensity events. However, when it rains heavily, sewer pipes can reach capacity and sewer backups may result. Damage claims are filed by the parties with sewer backup damage incurred on their property and, in some cases, the city will pay a damage claim amount to cover the amount of damage. The dataset used in this project contains sewer backups that caused a total of ;Meteorological, demographic, environmental and structural variables that explain the damage caused by those three storms are analyzed using a rare events logistic regression model. Sewer backups are rare events in Seattle since the highest claim-producing storm induced 147 claims in Seattle, a city with over 180,000 parcels. The model uses the claims from a particular storm and a random stratified citywide sample of parcels (stratified by neighborhood) to examine the explanatory variables that explain the occurrence of backups. A conditional backup probability is calculated for each sample parcel.;A spatial econometric model is used to measure the effect of explanatory variables that explain various levels of sewer backup damage while accounting for spatial effects of clustered claims. The results of the model are used to calculate potential damage for each sample parcel. The probability and potential damage calculations are multiplied together to produce an expected sewer backup damage (ESBD) amount for the sample parcels. These calculations were used to create three maps that represent probabilities of backups (conditional on the occurrence of a claim-producing storm), potential damage and ESBD.;These maps and the data that makes up the map can be used to prioritize preventative maintenance before a storm season. There are many other risks that face utility customers in Seattle but focusing on sewer backup risk allows for the application of two econometric models to better assess this specific risk. Such analysis has not been utilized to analyze the occurrence of sewer backups to date. Given the results of Salathe et al. (2010) and Zhu (2012) that suggest that higher frequency and higher intensity storms will affect the Puget Sound area, the accumulation of data and the use of the best information can efficiently mitigate damage caused by future storms.