A robust estimation of hedonic price models: least absolute deviations estimation

Abstract

Conventional parametric estimation of the hedonic price models is not robust to heteroscedastic and/or non-normal error structure. This paper applies least absolute deviations (LAD) estimation as a robust approach to estimating the hedonic price models, using the Korea housing markets data. The paper finds that LAD estimation produces more reasonable results and that it proves robust in a situation where other estimation results based on various functional form models produce inaccurate or misleading results.