Housing price determinants in Ecuador: a spatial hedonic analysis

• Date: 02 July 2024
• Pages: 1461-1487
• DOI: https://doi.org/10.1108/IJHMA-09-2023-0121
• Published date: 02 July 2024
• Subject Matter: Property management & built environment,Real estate & property,Housing markets
• Author: Felipe Miguel Valdez Gómez de la Torre,Xuwei Chen

Housing price determinants in

Ecuador: a spatial hedonic

analysis

Felipe Miguel Valdez G

omez de la Torre and Xuwei Chen

Department of Earth, Atmosphere and Environment, Northern Illinois University,

DeKalb, Illinois, USA

Abstract

Purpose –This paper aims to compare the eff‌iciency of spatial and nonspatial hedonic price models in capturing

housing submarkets dynamics for cities in developing countries. This study expects to contribute to a bett er

understanding of the housing price determinants from both nonspatial and spatial perspectives. In addition, t his

paper f‌ills a gap in the literature on the studyof h ousingprices from a spatial perspective in Latin American cities.

Design/methodology/approach –This study uses a comparative analysis between an ordinary least

squares regression and a geographical weighted regression,GWR. The study also assesses the performance

of two distinctdata sources: the city’s cadastralrecords and a real estate sales web portal.

Findings –The results suggest that compared to the traditional regression model, the spatial regression

models are more effective at capturing housing market variations on a f‌ine scale. Moreover, they reveal

interesting f‌indings on the spatial varying, sometimes contradictory effects of some housing attributes on

housing pricesin different areas of the city, suggesting the potential impact from segregation.

Research limitations/implications –The availability of data on housing prices and characteristics in

Latin American cities is fragmented and complex. The level of detail, granularity and coverage is not consistent

over time. For this reason, this study combines and compares data sets from off‌icial and unoff‌icial sources in an

effort to close this gap. Likewise, the socioeconomic variables that come from the censusmust be carefully analyzed,

knowing thehistorical context in which they were constructed, what they representand their interpretation.

Practical implications –This paper suggests that despite the improvementon the spatial models, the

selection of a specif‌ic one shouldalways be based on the diagnosis of it as it highly depends on the data used

and the objectivesof the study.

Originality/value –This study enriches the limitedbody of literature on spatial hedonic price modelsof

housing in Latin American cities. It also shed light on the importance of spatial approaches to identify

complex housingsubmarkets.

Keywords Latin America, Housing prices, Pricing model, GIS geographic information systems,

Geographically weighted regression, Spatial heterogeneity

Paper type Research paper

1. Introduction

Housing price is a primary constraint in household residential decisions, thus making it one of

the most important determinants of residential choices. Property value has been a major focus

in housing studies in the past f‌ive decades (Malpezzi, 2008;Sirmans et al., 2006;Zietz et al.,

2008). It is well understood that the determinants of housing prices, especially in urban

environments, are quite complex. Early studies have suggested that property prices are

This paper forms part of a special section “Spatial analysis and housing markets”, guest edited by

Koech Cheruiyot.

Housing price

determinants

in Ecuador

1461

Received18 September 2023

Revised22 December 2023

2 May2024

Accepted27 May 2024

InternationalJournal of Housing

Marketsand Analysis

Vol.17 No. 6, 2024

pp. 1461-1487

© Emerald Publishing Limited

1753-8270

DOI 10.1108/IJHMA-09-2023-0121

The current issue and full text archive of this journal is available on Emerald Insight at:

https://www.emerald.com/insight/1753-8270.htm

determined mainly by their physical characteristics (e.g. size, use, services), locations and other

external factors related to the nontangible values of the properties (Sirmans et al., 2006).

Since its introduction by Rosen (1974), the hedonic price model (HPM) has been the most

applied method for both estimating housing prices and identifying the contribution of the

elements related to housing prices. It is one of the earliest applications of multivariate statistical

techniques to housing price evaluation (Xiao, 2017). Within the HPM, a house is composed of

various attributes including structural characteristics and the surrounding spatial conditions.

The HPM allows to account for both a heterogeneous housing stock and the different ways

consumers value these characteristics (Malpezzi, 2008;Sirmans et al., 2005). Since its early

applications, the empirical results of the HPM suggested the existence of housing submarkets

based on the heterogeneity of the stock and preferences as well as the uniqueness of housing

location itself (Schnare and Struyk, 1976). However, most studies on housing prices are based

on data from cities in the developed countries (Abidoye and Chan, 2017;Chin and Chau, 2003).

Given the cultural, social, economic and morphological differences between cities in developed

and developing countries (Griff‌in and Ford, 1980), it is necessary to examine what variables are

inf‌luential in housing prices and how they affect housing prices in developing countries.

Under this context, this study uses both nonspatial and spatial regression models to examine

the factors associated with housing prices for the city of Quito, Ecuador. This study expects to

contribute to a better understanding of the housing price determinants from both nonspatial and

spatial perspectives. In addition, this paper f‌ills a gap in the literature on the study of housing

prices from a spatial perspective in Latin American cities. The purpose of this article, therefore, is

threefold. First, we use a traditional HPM –ordinary least squares model ( OLS) to identify which

factors inf‌luence housing prices in Quito on a global scale. Second, we compare the results of the

model based on two different data sets: the municipal housing appraisal data set and a real estate

data set. Lastly, we consider the impact of locations and examine the spatial varying effects of

those determinants on housing prices from a spatial perspective using geographical weighted

regression (GWR). We compare the results of two traditional HPMs with the GWR model. We

explore the existence of housing submarkets, where the coeff‌icients of the factors differ, within

the city. Findings from our study provide insights to the effectiveness of applying those models to

cities similar to Quito.

This paper is structured as follows. First, we provide a literature review of common

f‌indings on housing price determinants using HPMs, including their spatial variations and

case studies. Next, we use OLS and GWR models to analyze the determinants and their

spatial effects on housing prices in Quito. Then, we compare the results from those models

and discuss their relative effectiveness in capturing the characteristics of the housing price

and housing market in Quito. In the conclusion, we discuss the issues related to housing

price modeling and theapplicability of these models in Latin American cities.

2. Literature review

2.1 Hedonic price models (HPM models)

According to Rosen (1974), the value of a product is equal to the value assigned to each of its

attributes based on the utility perceived by consumers, which he calls implicit prices. The model

assumes a differentiated product market in which an equilibrium is reached when consumers are

willing to pay the implicit prices of the attributes offered by producers. In that respect, the model

allows the study of consumer preferences based on the implicit, or hedonic, prices of each of the

product attributes. It is then understood that market prices ref‌lect these preferences. Prices are

modeled from a vector of the prices of each of the product characteristics, in a linear regression.

Thus, the price of a house is determined by its structural characteristics (size, bathrooms,

materials) and by its location (specially accessibility to the central business district [CBD] as

IJHMA

17,6

1462