Automated Valuation Models Research Articles

Automated valuation models for real estate portfolios

PurposeAs regards the assessment of the market values of properties that compose real estate portfolios, the purpose of this paper is to propose and test an automated valuation model. In particular, the method defined allows for providing for objective, reliable and “quick” valuations of the assets in the phases of periodic reviews of the property values.Design/methodology/approachAiming at both predictive and interpretative purposes, the method, based on multi-objective genetic algorithms to search those model expressions that simultaneously maximize the accuracy of the data and the parsimony of the mathematical functions, is applied to a sample data of office properties characterized by medium and large size, located in the city of Milan (Italy) and sold in the period between 2004 and 2015.FindingsThe model obtained could be an integration of the canonical methodologies (market approach, income approach, cost approach) implemented in the assessment of the market values of properties, so as to provide an additional tool to verify the results. In particular, the inclusion of economic variables in the model is consistent with the need to reiterate the valuations, contextualizing them to the locational characteristics and to the current property cycle phase in the specific area.Practical implicationsThe model can be applied by all the operators involved in the periodic reviews of the values of property portfolios: from real estate funds’ insiders, in order to monitor the values obtained through the canonical approaches, to the public institutions, such as the revenue agencies, in order to ensure the fair payment of the taxes through the updating values of the properties according to the actual and current market trends.Originality/valueThe method proposed can be a valid support for all public and private entities that hold significant property assets and that, for various reasons (periodic reviews of the balance sheets, sales, enhancement, investment, etc.), require cyclical updated values of the properties. The automated valuation model developed can be used for the assessment of “comparison” values with the estimates values obtained by other assessment techniques, in order to ensure a further monitoring tool of the results from the subjects involved.

The future of the Australian valuation profession

PurposeValuers face significant challenges as processes become automated and the role evolves to data handling and processing. To survive and thrive, valuers must respond to a changing market. The purpose of this paper is to examine the issues, threats and challenges facing the Australian profession, though the issues are global.Design/methodology/approachThis qualitative research sought a deeper understanding of the threats, challenges and new areas of practice that Australian valuers were experiencing. A focus group approach was designed to collect data from practitioners in Sydney and Melbourne. The research aimed to identify new knowledge and skills for the future and emerging trends and practices.FindingsThe key issues, threats and challenges faced included increasing use of automated valuation models for low-risk residential valuations, valuers being unable to protect themselves against the banks, loss of control of the data and valuations. In total, 12 knowledge domains and skills required in the future were established and ten emerging trends and practices were identified.Research limitations/implicationsThe key limitations were that participants were from Melbourne and Sydney in Australia only and the focus is NSW and Victoria centric, although many participants have international work experience. There was an under representation of rural valuers, of small valuation firms, of young, recently joined or qualified valuers and females.Practical implicationsThe findings inform a manifesto for the future which sets out the practical implications for valuers and the professional body. This action plan sets the new knowledge domains, practices and trends that can be adopted by the profession and its members.Originality/valueThis is the original research and highlights some real threats, issues and challenges facing the Australian valuers. It complements work undertaken by legal and accounting professional bodies, which sense change affecting their membership and services. A manifesto for action has been outlined to address the changes that are coming and those already here.

Automating the land valuation process carried out in land consolidation schemes

Land consolidation is an effectual land management procedure aiming to sustainable development. A critical issue in land consolidation schemes is land valuation which is a type of mass appraisal that is generally carried out using an empirical process. As a result, the process is time consuming, costly and the outcomes are not adequately accurate, reliable and consistent. A solution to these problems is the employment of automated valuation models (AVMs). In this context, this paper presents the development, implementation and evaluation of two hedonic price models based on a linear and a non-linear function combined with a geographical information system (GIS), applied to a case study area in Cyprus. Models tested for quality assurance based on international standards. The evaluation based on international standards showed that the both models produced very good results and a small sample of land values provided by the Land Valuation Committee (LVC) is adequate to automate the land valuation process resulting in an adequate accuracy, reliability and consistency. Consequently, the AVM is highly efficient compared to conventional land valuation methods since it may considerably reduce the time and resources used and provide transparency because the process has been converted from empirical to systematic, analytical and standardized.

Real Estate and Land Property Automated Valuation Systems: A Taxonomy and Conceptual Model.

Automated valuation models have been in use at least for the last fifty years in both academia and practice, while a proper definition was coined only in the last decade. This could be mostly backed by the fact that research done on the automated valuation models is mostly empirical and evidently lacking a conceptual framework. This imply two-sided contribution of this paper, a taxonomy and a conceptual framework.

Determination of the parameters of automated valuation models for the hedonic property valuation of residential properties

PurposeThe development and testing of the hedonic methods for property valuation require statistical analysis and professional preparation of relevant databases. As a first step, the presumable relevant influencing variables (parameters) have to be determined. Previous studies have shown a large variety of parameters which overlap or deviate from each other. This study aims to collect, systematise and structure different parameters for the further development and testing of hedonic models.Design/methodology/approachThe study comprises a detailed research and deeper analysis of previous studies regarding the hedonic valuation (mainly for residential properties). Flanking areas of examination serve, if they are appropriately suitable, as supplements (e.g. performance analysis and regression). Parameters are extracted from a wide range of literature, compared and integrated into an overall presentation of the results.FindingsIn total, 407 parameters were extracted from previous studies on hedonic valuation and performance analysis. Because of various definitions of some parameters in the literature, the current paper combined them in one meaning to avoid misunderstandings in further analysis. Higher-level (global) and lower-level (specific) parameters are contained/described in the final list. The result of this study identifies up to five levels of parameters (within the relevant hierarchy).Research limitations/implicationsThe parameters have not yet been statistically tested. The relevance of individual parameters has to be tested with relevant corresponding databases and statistical methods (e.g. correlation and regression).Practical implicationsTo manage larger real estate portfolios, there is a need for regular property valuations. From this perspective, there is a great interest related to the optimisation of the valuation costs, valuation quality and valuation duration. Hedonic methods are considered as an efficient way of performing these valuation tasks. However, further suitable models and parameters are needed. The study describes parameters that can be appropriate for the development of relevant models and creates a structured parameters list, providing the technical basis for the latter. This structured parameter list is substantiated by the evaluation of the existing research.Social implicationsProperty values represent a significant asset for the national economy and for the individual wealth/welfare. The development of property value in a national economy is also relevant for politics, economy and society. The use of hedonic methods and the knowledge of important individual parameters can contribute towards assessing and substantiating the effect of political decisions on the value of real estate.Originality/valueFor the first time, a comprehensive and structured analysis on the value of the relevant parameters used in hedonic methods is performed. Thereby a large number of parameters were identified which question the stability of the results in respective individual studies. In addition, the newly developed hierarchy of parameters can serve as the basis for further research.

Big Data in Real Estate? From Manual Appraisal to Automated Valuation

Real estate is the third-largest asset class for institutional investors, but determining the value of commercial real estate assets remains elusively hard. In this article, the authors provide a practical application of big data by employing a unique set of data on U.S. multifamily assets, in combination with sophisticated modeling techniques, to develop an automated, machine-based valuation model for the commercial real estate sector. The authors find strong evidence of the superiority of automated valuation models over traditional appraisals: The absolute error of the automated model is 9%, which compares favorably against the accuracy of traditional appraisals, and the model can produce an instant value at every moment in time at a very low cost. The authors also provide evidence of the importance of using hyperlocal information on the location of an asset. The model developed in this article is directly applicable for real estate lenders and investors and has important implications for the traditional appraisal industry.

Collateral Misreporting in the RMBS Market

Securitized mortgage appraisals routinely target pre-specified valuations, 45% of purchase loan appraisals exactly equal purchase prices, and appraisals virtually never fall below purchase prices. As a result, appraisals exceed automated valuation model (AVM) valuations 60% of the time and are 5% higher than AVM valuations on average. High appraisals and indicators of appraisal targeting predict loan delinquency and RMBS losses and are priced at the loan level through higher interest rates, but have essentially no impact on RMBS pricing. Selection bias simulations and unfunded loan application appraisals indicate that high appraisals are intentional. The extent to which appraisals exceed AVM valuations varies across loan officers, mortgage brokers, and appraisers; and high appraisals are associated with more repeat business for appraisers, potentially incentivizing appraisers to inflate their appraisals.

Reliable region predictions for automated valuation models

Accurate property valuation is important for property purchasers, investors and for mortgage-providers to assess credit risk in the mortgage market. Automated valuation models (AVM) are being developed to provide cheap, objective valuations that allow dynamic updating of property values over the term of a mortgage. A useful feature of automated valuations is to provide a region of plausible price estimates for each individual property, rather than just a single point estimate. This would allow buyers and sellers to understand uncertainty on pricing individual properties and mortgage providers to include conservatism in their credit risk assessment. In this study, Conformal Predictors (CP) are used to provide such region predictions, whilst strictly controlling for predictive accuracy. We show how an AVM can be constructed using a CP, based on an underlying k-nearest neighbours approach. Time trend in property prices is dealt with by assuming a systematic effect over time and adjusting prices in the training data accordingly. The AVM is tested on a large data set of London property prices. Region predictions are shown to be reliable and the efficiency, ie region width, of property price predictions is investigated. In particular, a regression model is constructed to model the uncertainty in price prediction linked to property characteristics.

Indicators of Collateral Misreporting

Automated valuation models (AVMs) are widely used in the valuation of pools of residential mortgages. In this paper, we provide evidence that high pricing errors in human and automated valuation models as well as rejection of loans with relatively low appraisals may generate the appearance of appraisal bias even when the original appraisals are unbiased. We also present evidence that the estimated frequency of appraisal overstatement is quite sensitive to assumptions about loan denial rates due to low appraisals as well as the size of and the correlation between AVM and original appraisal pricing errors. This paper was accepted by Amit Seru, finance.

Effects of economic conditions and other factors on hotel sale prices

Purpose This study aims to examine previously identified hotel automated valuation models (AVMs) using a more extensive and updated sample of hotel sale transaction data, introduce economic conditions as a new determinant for hotel market value and test the moderating role of economic conditions on the relationship between the previously identified indicators and market value, operationalized as sale transaction price. Design/methodology/approach This study performs simple and multiple regression analyses to examine the proposed relationships. The sample period is from 2000 to 2012, because the economy prior to 2000 is likely to have been a less-relevant economic environment for the contemporary business world. Findings As hypothesized, there is an insignificant main effect of economic conditions on hotel sale prices. A moderating role of economic conditions on the relationship between average daily rate (and net operating income) and market value was found, while no moderating effect was found for occupancy rate and hotel size. Results also find certain support for AVMs previously presented in the literature, as described herein. Originality/value This study attempts to make contributions to the existing lodging and real estate literature by investigating a main effect of economic conditions on hotel sale prices and a moderating role of economic conditions on the relationship between various hotel performance indicators and hotel sale prices. In addition, this study operationalizes a relatively extensive (large base sample size of 2,441 hotel sale transactions) and recent (since 2000) database compared to previous studies.

A spatially based artificial neural network mass valuation model for land consolidation

Land consolidation, which aims to promote sustainable development of rural areas, involves the reorganization of space through land reallocation, both in terms of ownership and land parcel boundaries. Land reallocation, which is the core part of such schemes, is based on land values because each landowner is entitled to receive a property with approximately the same land value after land consolidation. Therefore, land value, which in the case of Cyprus is the market value, is a critical parameter, and hence it should be reliable, accurate, and fairly valued. However, the conventional land valuation process has some weaknesses. It is carried out manually and empirically by a five-member Land Valuation Committee, which visits every unique parcel in the consolidated area to assign a market value. As a result, it is time consuming and hence costly. Moreover, the outcomes can be inconsistent across valuators for whom, in the case of such a mass appraisal procedure, it is hard to analytically calculate the scores for a series of land valuation factors and compare all of these for hundreds of land parcels using a manual process. A solution to these shortcomings is the use of automated valuation models. In this context, this paper presents the development, implementation, and evaluation of an artificial neural network automated valuation model combined with a geographical information system applied in a land consolidation case study area in Cyprus. The model has been tested for quality assurance based on international standards. The evaluation showed that a sample of 15% of the selected land parcel values provided by the Land Valuation Committee is adequate for appraising the land values of all parcels in the land consolidation area with a high or acceptable accuracy, reliability, and consistency. Consequently, the automated valuation model is highly efficient compared to the conventional land valuation method since it may reduce time and resources used by up to 80%. Although the new process is based partly on the Land Valuation Committee sample, which inherently carries inconsistencies, it is systematic, analytical, and standardized, hence enhancing transparency. The comparison of artificial neural networks with similar linear and nonlinear models applied to the same case study area showed that it is capable of producing better results than the former and similar outcomes to the latter.

Indicators of Collateral Misreporting

Recent studies use retrospective automated valuation models (AVMs) to investigate whether appraisals of collateral values associated with securitized residential mortgage loans were systematically inflated during the recent housing boom. In this paper, we provide evidence that high AVM pricing errors as well as selection bias in samples of completed loans, which are conditioned on relatively high appraisals, may generate the appearance of appraisal bias even when the original appraisals are unbiased estimates of the underlying value of the collateral. We present evidence that the estimated frequency of appraisal overstatement is quite sensitive to assumptions concerning loan denial rates as well as the magnitude of and the correlation between AVM and original appraisal pricing errors.

Understanding House Price Index Revisions

Residential house price indexes (HPI) are used for a large variety of macroeconomic and microeconomic research and policy purposes, as well as for automated valuation models. As is well known, these indexes are subject to substantial revisions in the months following the initial release, both because transaction data can be slow to come in, and as a consequence of the repeat sales methodology, which interpolates the effect of sales over the entire period since the house last changed hands. We study the properties of the revisions to the CoreLogic House Price Index. This index is used both by researchers and in the Financial Accounts of the United States to compute the value of residential real estate. We show that the magnitude of revisions to this index can be significant: At the national level, the ratio of standard deviation of monthly revisions to the growth rate of the index, relative to the standard deviation of the growth rate in the index, is 29%, which is comparable to the relative ratio for other macroeconomic series. The revisions are also economically significant and impact measures used by policymakers: Revisions over the first 12 releases of the index reduce estimates of the fraction of borrowers nationwide with negative equity by 4.3%, corresponding to 423,000 households. Lastly, we find that revisions are ex-ante predictable: Both past revisions and past house price appreciation are negatively correlated with future revisions.

Automated valuation model for livestock appraisal in loaning practice

Actualization of loan security (mortgage) value is of major importance in Hungarian loaning practice. Due to the recession in economics, the value of agricultural portfolio of banks has decreased a great deal, though not to such a great extent as other branches of the economy. Depreciation of estate stock is compensated with additional collateral security. Besides other stock, often temporarily and out of necessity, livestock is presented as additional collateral security. From the loaners’ point of view, however, the registered inventory value does not guarantee security. The authors have set up an appraisal method giving professional guidance through automated valuation as to how dairy stock can be used as mortgage for loan security. Hereby we are to present the details of both the theory and the methodology of a model that is appropriate for the valuation of dairy livestock on an MS Excel basis. Thus, the process is fast and has more prospects for all parties in the loaning or leasing business. The method involves the features of livestock technology, the expected realized profit, and breed stock value. By the implementation of this method, the loaners can calculate the value of loan recovery (loan to value) with acceptable security.

Automated valuation modelling: a specification exercise

Market value predictions for residential properties are important for investment decisions and the risk management of households, banks and real estate developers. The increased access to market data has spurred the development and application of Automated Valuation Models (AVMs), which can provide appraisals at low cost. We discuss the stages involved when developing an AVM. By reflecting on our experience with md*immo, an AVM from Berlin, Germany, our paper contributes to an area that has not received much attention in the academic literature. In addition to discussing the main stages of AVM development, we examine empirically the statistical model development and validation step. We find that automated outlier removal is important and that a log model performs best, but only if it accounts for the retransformation problem and heteroscedasticity.

A Reexamination of Current Hotel Valuation Techniques – Which Approach is More Realistic?

This study revisits the issue of accuracy in contemporary hotel valuation. Along with the hotel valuation techniques used by Rushmore (1992) and Chen and Kim (2010), this study uses the cost approach and the automated valuation model (AVM) in its examination of contemporary hotel valuation techniques. Fourteen randomly selected hotel firms are analyzed using nine valuation approaches. The valuation results are then compared to the market values of these firms to assess which technique provides the most robust and supportable estimate. Research results reveal that, at least for the analyzed sample, the discounted cash flow (DCF) technique provides the most realistic estimate of a hotel firm's value. Results also show that the valuation estimate of AVM is significantly different from both Band of Investment methods. As such, the process of valuing hotel properties is better understood.

Latest Developments in Mortgage Analyticsfor Investors

Raj Dosaj explained various factors that contribute to the accuracy of home price indexes, such as geo-granularity, attributes of the McDash database, and the Lender Processing Services (LPS) Credit Bureau Module. Michael Bradley explained a set of disposition models developed by CoreLogic that estimate distressed sale values at various stages of the liquidation process that he believes are more accurate, less expensive, and can be obtained more quickly than broker price opinion (BPO) valuations. Dan Abraham discussed the use of Lewtan’s address-specific rental automated valuation models (AVMs) and how those models are reflected in statistics at the neighborhood and ZIP Code levels. Ron D’Vari discussed challenges with loan-level models for U.S. RMBS (residential mortgage-backed securities), including too many parameters and not enough time to validate interdependencies and the dominance of macro factors, such as interest rates and HPIs (house price indexes), and how risk-based cohorting provides a solution to incorporate loan-level performance. Brian Grow explained Morningstar’s approach to rating structured securities, in which issuers pay for initial ratings, investors pay for surveillance, a proprietary loan-level credit model is used, and all rated securities are re-rated monthly. <b>TOPICS:</b>MBS and residential mortgage loans, factor-based models

A Space-Time Model for Computer Assisted Mass Appraisal

An Automated Valuation Model (AVM) that seeks to attain predictive accuracy must take into account both spatial and temporal effects in the real estate market. A model structure that contains neither explicit spatial nor temporal variables is calibrated by a method that recognizes these variations in is calibration architecture. The method is conceptually similar to Geographically Weighted Regression (GWR) except that it extends into the temporal domain. The methodology is explained and results provided illustrating spatio-temporal variations in value.

Technologies To Improve The Decision-Making Process Of Real Estate Appraisers: XML, Intelligent Agents, Avms, And Web Services

&lt;p class="MsoBlockText" style="margin: 0in 0.5in 0pt 1in;"&gt;&lt;span style="font-style: normal; mso-bidi-font-style: italic;"&gt;&lt;span style="font-size: x-small;"&gt;&lt;span style="font-family: Times New Roman;"&gt;Pressure to expedite mortgage originations has prompted the need to accelerate the appraisal process for residential real estate.&lt;span style="mso-spacerun: yes;"&gt;&amp;nbsp; &lt;/span&gt;The appraisal aspect of a mortgage origination is one area where decision support can play a major role in redefining the process. This study presents propositions that incorporate new technologies into the decision-making process of a residential real estate appraiser.&lt;span style="mso-spacerun: yes;"&gt;&amp;nbsp; &lt;/span&gt;The four propositions deal with intelligent agents, XML, automated valuation models (AVMs), and web services. Intelligent agents have the ability to redefine how appraisal firms create, maintain, and update their databases.&lt;span style="mso-spacerun: yes;"&gt;&amp;nbsp; &lt;/span&gt;XML offers enhanced data handling capabilities, while automated valuation models offer the ability to drastically reduce the time required to render an estimate of value. These two technologies can be combined with web services to make the real estate appraisal application available anywhere and anytime of the day.&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;

Of Mice and Men

The rise in the use of Automated Valuation Models (AVMs) poses questions about the long-term survival of the property valuation profession. AVMs provide quick and cheaper valuations which in this very competitive market-place put them at a distinct advantage over traditional approaches. This article engages with the normative question of whether AVMs could completely replace traditional valuers and whether it is appropriate to do so. Going to ‘first principles’, this paper builds an alternative conception of market value in terms of which both AVMs and traditional valuations are argued to be based on false ontological assumptions. The paper locates both AVMs and traditional valuations in the social processes in which market value is created. The paper concludes that there are no theoretical or practical reasons why AVMs should not completely replace traditional valuers.