No large-scale empirical study on condemnation compensation has been done in the past 30 years. Several state legislatures, in response to Kelo v. City of New London, have changed laws to increase condemnation compensation, despite the lack of empirical grounds. To fill in the empirical gap, I use hedonic regression models and about 80,000 sales to estimate the fair market value (FMV) of 430 condemned properties whose owners reached compensation settlements with the condemnor, New York City, between 1990 and 2002. More than 50 percent of these condemnees were compensated with less than FMV, about 40 percent received more than FMV, and less than 10 percent received FMV. Owners of residential properties and nonresidential properties alike often received extreme compensations that are less than 50 percent or more than 150 percent of FMV. Extreme compensation results from bias-prone and inaccurate appraisal methods. Using the available data, I find that compensation level does not correlate with any factor. (c) 2010 by The University of Chicago. All rights reserved..