Abstract
This paper is to analyze office rent determinants using spatial autocorrelation, especially focusing on sub-market classified by office building grades and regions. This paper uses office rent data in Seoul at third quarter of 2009. Most researchers has recognized that office rents has been spatially influenced among themselves. There has, however, been little research for sub-market considering spatial autocorrelation by grades and regions, if any. Spatial models outperforms OLS model on the grounds of adjusted R-squares. This study shows that office rents are spatially correlated in the market, but not in prime office market. In A class office market, weak spatial correlation is observed. There is also spatial correlation among regional office markets. The fact means that autocorrelation must be considered for more correct analysis, and different factors must be also considered for analyzing office sub-market by grades and regions.