Value Drivers and Valuation Algorithms of Residential Investments

Abstract

This paper presents the results of two valuation functions of residential investments properties in Berlin, Germany and thus the influencing factors on the price of these properties. With their assistance, house prices are predicted, which take into account the heterogeneous features of residential properties. The estimation procedure uses multiple regression models, so-called hedonic models. These hedonic models evaluate properties based on their intrinsic values and determine the overall value from their distinguishing characteristics. The focus is on multi-dwelling units [i.e. ‘houses in multiple occupation [HMOs], multiple-family dwellings, etc.] and rented condominiums, which is to say classical investment properties that are typically appraised using the income approach or the discounted cash flow method [DCF method]. Sales comparison approaches such as the hedonic method are considered to be more appropriate for property valuation due to their closeness to the market and reduced possibility of subjective approaches. With digitisation, automated valuation procedures have become more important and also enabled alternative methods that require less costs, effort and subjectivity for the appraiser. In this Paper, the question is addressed as to whether or not a valuation algorithm for estimating market prices by means of statistical methods is also applicable for real estate investments.