Correcting <i>HAZUS</i> General Building Stock Structural Replacement Cost Data for Single-Family Residences

Abstract

Alternative approaches are taken to correct HAZUS General Building Stock (HGBS) replacement cost data for single-family residences in Sarpy County, Nebraska, which are 92% higher than tax assessor estimates of depreciated structural replacement values (DSRVs). First, HGBS costs were reduced by 92% across all census blocks, resulting in mean block-level HGBS–assessor differences of −31% with a standard deviation of 144%. Second, DSRVs were regressed against non-cost-related HGBS data at the block level via six alternative model specifications, resulting in R2 values of 0.59–0.68. The preferred model based on HGBS square footage and other census-based data resulted in predicted HGBS replacement costs on average being 29% higher than DSRVs, although the standard deviation of the differences was 380%. Finally, regressing DSRVs against tax assessor building square footage data resulted in a R2 value of 0.96 and an underprediction of actual DSRVs across blocks by only 12%, with a correspondingly low standard deviation of 48%.