Edition 2020 Ninth edition
EVIP 7 Advanced Statistical Models 1
Download 1.74 Mb. Pdf ko'rish
|
a6048c931cdc93 TEGOVA EVS 2020 digital
EVIP 7 Advanced Statistical Models
1. Definition 2. Introduction 3. Commentary 4. European Banking Authority criteria for advanced statistical models for valuation and EVS commentary 5. The valuer's use of statistical tools 288 IV. - EVIP 7: Advanced Statistical Models European Valuation Standards 2020 EVS adheres to the European Banking Authority's Guidelines. Relevant excerpts from the EBA Final Report — Guidelines on loan origination and monitoring; 7 Valuation of immovable and movable property; 7.1 Valuation at the point of origination — 29 May 2020 209. At the point of origination, institutions should ensure that the value of all immov- able property collateral for loans to consumers and micro, small, medium-sized and large enterprises is assessed by an internal or external valuer using full visit with internal and external assessment of the property. 210. As a derogation from paragraph 209, for the purposes of a valuation of residen- tial real estate in well-developed and mature property markets, the value may be assessed by means of a desktop valuation, carried out by an internal or ex- ternal valuer and supported by advanced statistical models. The valuer remains responsible for the valuation, while the advanced statistical models should be used as supporting tools, meeting the conditions set out in Section 7.4, and in- cluding a confidence measure to indicate the robustness of the value proposal and other relevant property-specific information. In this case, the value propos- al should be assessed, reviewed and approved by the internal or external valuer, who should understand all inputs and assumptions considered in the model. If the confidence measure in the supporting advanced statistical model indicates low robustness, and/or other property-specific information gives rise to uncertainty about the value proposal, the valuer should choose a valuation method other than desktop valuation. […] 225. When the conditions for a review in accordance with Article 208(3)(b) of Regula- tion (EU) No 575/2013 are met, institutions should update the value of the immov- able property collateral by means of a revaluation carried out by a valuer who is potentially supported by appropriate advanced statistical models that meet the conditions set out in Section 7.4 and account for individual characteristics of the property and geographical area. Institutions should not use these models as the sole means of the revaluation. European Valuation Standards 2020 IV. - EVIP 7: Advanced Statistical Models 289 1. Definition 1.1. Advanced statistical models are the most modern and sophisticated automat- ed valuation models (AVMs), statistically-based computer programmes which use property information to generate property-related values or suggested values. 1.2. A statistical method of valuation seeks to arrive at the value of a property directly through the application of a mathematical algorithm to a data base of transaction prices and property characteristics whilst omitting a professional valuer's quali- tative assessment of the value. Hence, in the context of real estate valuation, an AVM is simply the mechanical application of an algorithmic procedure to the data presented to it, without taking into account a valuer's opinion as to other relevant information, without a physical inspection of the property and ignoring the weighting a valuer would place on the relevant information. 1.3. A stand-alone AVM estimate is not a valuation and cannot give a Market Value as defined by EVS 1. It can be used by qualified valuers as one input in their compre- hensive analysis of the market provided they are satisfied with their knowledge of the AVM's input data and model. 2. Introduction 2.1. Used as a tool for valuing properties quickly and at low cost, AVMs have intrin- sic limitations: • They can only work from the figures for the transactions reported without in- formation or insight into their context; • There is no inspection of the property. They must rely on an implicit assumption that the property is in marketable condition; • Limited coverage of relevant data in some areas; • Limited coverage of relevant recent data for some properties; • A limited ability to reflect any unique characteristics of the property; • They are unlikely to be relevant for specialist properties. 290 IV. - EVIP 7: Advanced Statistical Models European Valuation Standards 2020 There is little hard impartial evidence in the public domain or independent evalua- tion of the accuracy of AVMs for individual properties as European AVM manufac- turers are reluctant to release details and the underlying data or the construction of the algorithms are not made available for analysis 1 . 2.2. Those limitations have less potential negative consequences concerning the standard uses of AVMs in considering properties collectively, as for: • Banks identifying property that needs revaluation; • In-arrears assessment in banks; • Identification of fraudulent activity in banks; • Full valuation audits in banks; • Determining capital adequacy ratios in banks; • Mark-to-Market portfolio of properties in banks; • Mass Appraisal for local taxes by government; • Estimating relocation compensation by government; • Cost/Benefit analysis for potential public expenditure. 2.3. In recent years, AVMs have become much more widespread as a result of a drive towards rationalisation in the property and financial sectors. Abusive use of AVMs was one of the reasons for the financial crisis, as was stated in the final report of the parliamentary committee of inquiry of 27 January 2016 regarding the Irish banking crisis Download 1.74 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling