Resampling Techniques for Real Estate Appraisals: Testing the Bootstrap Approach

Vincenzo Del Giudice,Pierfrancesco De Paola,Francesca Salvo

doi:10.3390/su10093085

Vincenzo Del Giudice, Pierfrancesco De Paola + Show 1 more

Open Access

https://doi.org/10.3390/su10093085

Copy DOI

Abstract

Applied to real estate markets analysis, the resampling methods aim to contribute to the knowledge growth of real estate market dynamics, overcoming the issues related to the data scarcity and operational limits of traditional statistical theory. Among resampling methods, the Bootstrap technique appears to be the most suitable for the interpretation of real estate phenomena. In this study, for residential properties located in Cosenza (Calabria Region, Italy), a Bootstrap approach has been used in order to determine the marginal prices of the real estate characteristics detected, comparing the results with those obtainable with a traditional Multiple Regression Analysis.

Highlights

In the presence of a sufficient amount of real estate data, the traditional statistical theory postulates a normal distribution for the real estate prices, requiring the adoption of specific statistical measures applicable to the data population
From a general and statistical point of view, for Multiple Regression Analysis (MRA), it must be considered that the distribution of the independent variables is irrelevant if the assumptions of homoscedasticity and normality are fulfilled for the residuals
The marginal prices calculated by applying the MRA to the original real estate sample, as well as the marginal prices obtained according to the Bootstrap procedure, show slight divergences

Summary

Introduction

In the presence of a sufficient amount of real estate data, the traditional statistical theory postulates a normal distribution for the real estate prices, requiring the adoption of specific statistical measures applicable to the data population (e.g., mean, median, variance, standard error, standard deviation, etc.). In the real estate field, there is usually a low amount of detectable real estate data: this circumstance, together with the stratification processes of real estate markets, are conditions that do not satisfy the postulate of a normal distribution of the observed real estate prices [1,2,3,4,5,6,7,8,9,10,11,12]. From a general and statistical point of view, for Multiple Regression Analysis (MRA), it must be considered that the distribution of the independent variables (real estate characteristics) is irrelevant if the assumptions of homoscedasticity and normality are fulfilled for the residuals. For a real effect on the data, dependent variables (e.g., real estate prices) might often follow a Gaussian distribution, but this does not in any way imply that Gaussianity of dependent variables is necessary for the MRA to work [13]. It is clear that this hypothesis disregards the possibility of “deviation” (or variability) of the estimation function with respect to the statistical distribution of the comparable properties universe [14,15]

Methods

Results

Conclusion