Acceso usuarios
Ucrania
The article behaves to the property economic measurements implementation using the Comparative Sales independent valuation approach. On example from the real commercial real estate evaluation practice the main methodological principles of valuation object spatial localization characteristics adjustment are considered. According to the described methodology, localization adjustment coefficient is determined by calculation method on the basis of market data cross-correlation regressive analysis. A basic hypothesis is a statement that the relation of valuation object single unit value index to the same of comparable object is determined by its model values relation in the mathematical model of statistical relationship between object single unit value index and its three settlements characteristics: population number; distance to the regional center; area (territory within the settlement boundaries).
The research is grounded on mathematical simulation and mathematical statistic quantitative methods. The methodology of adjustment coefficients on investigated price-forming factors definition is based on nonlinear cross-correlation regressive analysis of market data research. This mathematical model is experimentally set by local market data research for the exactly similar real estate objects on the valuation date. It is set that there is observed different statistical relationship level between some of objects price-forming factors and its single square value indexes. The closest statistical relationship exists between settlements population number and single indicator of similar property situated in other compared settlements.
It is shown that taking into account some recommended braking coefficients for regressive curve, are published in professional literature, is inadvisable, because it increases the result error is got. Certainly the regression curve characteristics of object spatial localization price-forming factors must be taken into account at adjustment coefficient determination procedure. It is well-proven that methodically correct result of object localization adjustment procedure implementation can be provided only in the case of local market situation research data applying, with determination of the nonlinear regression function characteristics for statistical dependence of single square value index from the object settlements population number.
Research is described gives an opportunity to decrease evaluation result uncertainty through the use of new offered approach to mathematical model characteristics definition. The main result of researches described is a possibility to obtain appraising/valuation results with the higher reliability and better accuracy. Researches results are the objective confirmation of the fact, that nowadays methodical base of independent valuation is not able to provide the higher level of this class evaluation objects accuracy results. It does not depend only from an individual appraiser or concrete evaluation company, but, firstly, from unreliable arbitrarily chosen by appraisers adjustments - that usually are "expertly" determined, based on appraiser's own ideas about the dependence of real estate prices on the settlements characteristics. This elementary way of taking these characteristics into account may be a source of result additional errors and its uncertainty level increasing.
Future investigations in this direction may deals with the consideration and analysis of other types nonlinear functions application possibilities, that is approximate the regression curve of statistical interdependence between the object single value index and its spatial localization characteristics. The quantitative indexes of absolute and relative methodological errors also may be determined and analyzed in detail in future researches. The importance of those researches for the further development of the independent valuation metrological-information paradigm are confirmed. Practical recommendations for the evaluation results accuracy and reliability increasing are formulated.
