Printer Friendly

Taking a Look at AVM Challenges and Innovations.

Advances in Automated Valuation Modeling: AVM After the Non-Agency Mortgage Crisis Maurizio d'Amato and Tom Kauko, Editors

Published by Springer International Publishing AG, Germany, 2017, 418 pages; $179 hardcover, $139 ebook

Automated valuation models (AVMs), regression analysis, and advanced statistical analysis are most often considered in terms of their usage in the United States, but it is increasingly clear that their usage is far more pervasive and ultimately equally relevant in economies throughout the world. The editors of Advances in Automated Valuation Modeling: AVM After the Non-Agency Mortgage Crisis have assembled a group of authors from around the world to address various questions and demonstrate the research and practical implementations of AVM advances that are occurring across the globe.

The editors' previous book, Mass Appraisal Methods: An International Perspective for Property Valuers (Wiley-Blackwell, United Kingdom 2008), provided a discussion of mass appraisal techniques being developed and used internationally. That book also considered in-depth research into state-of-the-art developments that were likely to permeate the industry over the next decade.

Advances in Automated Valuation Modeling: AVM After the Non-Agency Mortgage Crisis is the culmination of the earlier text and demonstrates the advances and ongoing research that is occurring.

In their latest book, the editors have focused their efforts on providing an overview of mass appraisal methods that meet the needs of their local economic environments. The text is a compilation of articles by various authors with a range of international backgrounds. This international perspective effectively demonstrates that advances in technique and methodology can be applied to situations regardless of geography.

Advances in Automated Valuation Modeling is organized into five interrelated themes or parts: AVMs, mortgage crises, and valuation in person; mortgage crises and experiences in AVMs; AVM methodological challenges dealing with spatial issues; AVM methodological challenges and nondeterministic modeling; and AVM methodological challenges with inputs and models. Each section provides a global perspective on the use, experience, and future of AVMs.

The book begins with an examination of the theoretical foundations of AVMs, moves to an analysis and case studies in institutional contexts, and then moves on to discuss the practical usage in various markets around the world.

The discussion of AVMs in various markets includes an examination of diagnosed system failures, best practices, a comparison of theoretical and practical applications, and the types of challenges that lie ahead in promoting the continued and expanded usage of AVMs around the world. The examination of emerging problems related to AVMs begins with an investigation of some of the factors that led to the fiscal crisis of 2008. It specifically addresses the role of the appraiser and the role of AVMs in the analysis of real estate valuation. The crisis is examined from differing points of view, and the book integrates the current methodology by examining the root of the crisis in both academic and practical usage.

Of particular note is the linkage that is provided in part one between neoclassical economic theory and mathematical and statistical modeling. In an excellent overview that provides a critical understanding of why statistical modeling in real estate differs from the traditional scientific method, the book examines the construction of AVMs and their linkages to classical economics. The text notes:
   AVMs are based on an inductive-statistical model
   rather than a deductive-nomological model. This means
   that while deductive models are based on scientific
   principles and are appropriate for phenomena in the
   natural sciences (because these tend to represent exact
   scientific relationships), inductive-statistical models
   such as AVMs are probabilistic, meaning that they
   utilize probability, which is less exact, and relies on the
   concept of measuring social phenomena, which are far
   less exact. In other words, AVMs are based on the less
   exact methodology of prediction. This has implications
   for their predictive power, noting that they are less
   predictive than deductive models. Only incomplete or
   partial information is known, and for this reason, probability
   is required. Herein lies part of the explanation
   for the observed inaccuracies in AVMs. This also
   explains why prediction in the economics discipline
   (and the social sciences generally) has not, and possibly
   could not, match the successes found in the natural
   sciences, (page 48)


Another area of strength in the opening chapters of the book is the comparison of appraisers, AVMs, the role of the appraiser, and the role of the AVMs. In examining the differences between appraisers and AVMs, the book notes:
   In the aftermath of the 2008 non-agency mortgage crisis,
   questions have been raised about the role and
   potential culpability of the valuation (or appraisal) profession
   in precipitating it, or at best, in the collective
   failure to avert it. The role of AVMs has been the subject
   of critical commentary, with some going so far as to
   attribute the crisis to their widespread usage in the
   home loan origination process. It is a well-known fact
   that the period leading up to 2008 saw AVMs taking an
   increasingly larger share of valuations for mortgage purposes,
   at the expense of human appraisers, (page 53)


While many books and articles have examined the role of market participants in assessing blame for the budgetary crisis of 2008, this book is one of the few to present a measured view of the role of valuation in the crisis and to provide answers to some of the questions of the role of AVMs in the mortgage crisis and the relationship between the sophisticated automated valuation process and the valuations provided by appraisers. Specifically, the text notes:
   Explaining the culpability of appraisers for the budgetary
   crisis solely, or mainly, in terms of unethical and unprofessional
   behavior is, however, neither an adequate
   account nor particularly helpful. It does not account, for
   instance, for the behavior of ethical and honest appraisers
   doing valuations in the heady days of the property
   boom. In an environment of continually rising prices, did
   these honest appraisers interpret the market in a manner
   that fueled further rises? Could they have done otherwise,
   considering the theoretical at their disposal? The
   unfair assumption that unethical conduct fueled the
   problem may serve to detract attention from the insidious
   role played by conventional valuation theory and
   methodology in which AVMs are based. Links can be
   made between neoclassical economics and a strong
   argument has been made that the contribution that
   AVMs may have played towards the crisis stems from an
   underlying flaw in that AVMs are ill-equipped to handle
   market crises and instability, and to a methodological
   orientation that precludes or eschews, the possibility of
   appraiser influence in the pricing of property, (page 53)


The next sections of Advances in Automated Valuation Modeling are comprised of contributions from academics and practitioners throughout Europe, including Italy, Germany, and Turkey. The relationship between the financial and real estate sectors in emerging markets such as China is also examined, and the particular challenges in such markets are presented in a manner that provides insight into more sophisticated markets such as the United States.

The core of the book highlights some of the methodological challenges inherent in the process of improving AVMs, and specifically discusses concerns in the theoretical and empirical trends on the topics of spatial variables, the use of multilevel modeling to improve deterministic modeling, the role of nondeterministic reasoning, and finally the challenges of data paucity in valuation models and their inputs.

An examination of different cutting-edge AVM approaches constitutes the next section of the book, as topics such as utilizing fuzzy logic in an integrated multiple regression hedonic price model are examined. In this examination, the analysis permits the analyst to take into account the continuity of variables, and in this way, is appropriate as a means of improving existing AVM modeling techniques.

The lack of data in many markets is also given significant consideration, as the lack or unavailability of data can constrain the use of AVMs in many geographies. The conclusions in this section examine the potential of using rough set theory as a method of overcoming the challenges for both residential and commercial uses, by examining local market dynamics to account for the specific nature of market information that may be available. The questions raised for readers in this section are related to the relationship between modeling and observations to specific contexts where data are not abundant or may present a relationship between dependent and independent variables that may be modeled using different approaches.

The use of multiple authors brings a wide range of perspectives from across the globe. Similar to their earlier book, the editors are able to weave a coherent text that examines many related subjects, while at the same time providing a historic foundation and viewpoints on the potential for future growth. The book will appeal to practitioners, AVM developers, and those with an interest in how the evolution of valuation techniques and technology continues to bring both opportunity and risk mitigation, while at the same time contextualizing the role of traditional appraisal techniques. Given that the case studies and techniques that are presented concern both commercial properties and residential properties, this book will be of interest to a broad constituency of appraisers and those involved in financial services throughout the industry. Ultimately, Advances in Automated Valuation Modeling: AVM After the Non-Agency Mortgage Crisis provides a valuable addition to the body of knowledge and highlights the significant research work being performed to innovate globally in the context of creating more accurate and meaningful valuation techniques.

About the Reviewer

Mark R. Linne, MAI, SRA, AI-GRS, is chief executive officer of ValueScape, LLC, a San Diego-based data science technology company.

by Mark R. Linne, MAI, SRA, AI-GRS

To obtain books reviewed in The Appraisal Journal, please contact your local or online bookseller.
COPYRIGHT 2017 The Appraisal Institute
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Advances in Automated Valuation Modeling: AVM After the Non-Agency Mortgage Crisis
Author:Linne, Mark R.
Publication:Appraisal Journal
Article Type:Book review
Date:Jun 22, 2017
Words:1603
Previous Article:Latest data, latest literature.
Next Article:Water Impacts, Data Analysis.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters