표본선택편의를 고려한 오피스 매매가격 결정요인 분석 및 매매가격지수 산정

Abstract

Price determinants and indices that based on the sample of transacted offices may not accurately represent a population due to sample selection bias. We analyse sample selection bias in estimating parameters of a hedonic price equation in the Seoul metropolitan office market. We use Heckman two-stage procedure(Heckman), maximum likelihood estimation(ML) and ordinary least square estimation(OLS) as analysis methods. We find the impact of sample selection bias by comparing results of Heckman and ML with OLS. Some Independent variables like building height, age, lot area etc are turned out statistically significant in Heckman and ML but some of them are not significant in OLS. We calculate office price indices based on Heckman, ML and OLS. Consequently, Heckman and ML price indices are a little different from OLS price index but not strongly.