The hedonic house price index for Poland modelling on NBP BaRN data. Narodowy Bank Polski International Workshop, Zalesie Górne, November 2013

Similar documents
Krzysztof Olszewski, Krystyna Gałaszewska, Andrzej Jakubowski, Robert Leszczyński and Hanna Żywiecka

Using Hedonics to Create Land and Structure Price Indexes for the Ottawa Condominium Market

Modelling a hedonic index for commercial properties in Berlin

Metro Boston Perfect Fit Parking Initiative

Technical Description of the Freddie Mac House Price Index

A. K. Alexandridis University of Kent. D. Karlis Athens University of Economics and Business. D. Papastamos Eurobank Property Services S.A.

Review of the Prices of Rents and Owner-occupied Houses in Japan

Regional Housing Trends

An Assessment of Current House Price Developments in Germany 1

Meeting of Group of Experts on CPI 30 May 1 June 2012

TECHNICAL ASSISTANCE REPORT RESIDENTIAL PROPERTY PRICE STATISTICS CAPACITY DEVELOPMENT MISSION. Copies of this report are available to the public from

Modeling Farmland Conversion with New GIS Data

Calculating a constant quality price index for the stock of residential housing

86 years in the making Caspar G Haas 1922 Sales Prices as a Basis for Estimating Farmland Value

Automated Valuation Model

Housing Price Prediction Using Search Engine Query Data. Qian Dong Research Institute of Statistical Sciences of NBS Oct. 29, 2014

Estimating User Accessibility Benefits with a Housing Sales Hedonic Model

Hedonic Pricing Model Open Space and Residential Property Values

Hedonic analysis of office and retail rents in the three major cities in Poland

Commercial Property Price Indexes and the System of National Accounts

Frequently Asked Questions: Residential Property Price Index

The purpose of the appraisal was to determine the value of this six that is located in the Town of St. Mary s.

PROPERTY BAROMETER FNB House Price Index Early signs of the positive national sentiment shift impacting on national house price trends

MONTHLY HOUSE PRICE INDEX REPORT

Estimating National Levels of Home Improvement and Repair Spending by Rental Property Owners

14 N O V E M B E R

Housing Price Index, base 2007 Methodological preview

Description of IHS Hedonic Data Set and Model Developed for PUMA Area Price Index

Re-sales Analyses - Lansink and MPAC

An Assessment of Recent Increases of House Prices in Austria through the Lens of Fundamentals

Analysis on Natural Vacancy Rate for Rental Apartment in Tokyo s 23 Wards Excluding the Bias from Newly Constructed Units using TAS Vacancy Index

A Real-Option Based Dynamic Model to Simulate Real Estate Developer Behavior

Messung der Preise Schwerin, 16 June 2015 Page 1

The Corner House and Relative Property Values

Real Estate Price Index Measurement: Availability, Importance, and New Developments

An Introduction to RPX INTRODUCTION

How should we measure residential property prices to inform policy makers?

University of Zürich, Switzerland

Online Appendix "The Housing Market(s) of San Diego"

The impact of the global financial crisis on selected aspects of the local residential property market in Poland

Price Indices: What is Their Value?

WORKING PAPER N MEASURING AMERICAN RENTS: A REVISIONIST HISTORY

Sponsored by a Grant TÁMOP /2/A/KMR Course Material Developed by Department of Economics, Faculty of Social Sciences, Eötvös Loránd

A New Approach for Constructing Home Price Indices in China: The Pseudo Repeat Sales Model

Residential Property Index Series. August 2017

Commercial Property Price Indices for Greece

1. There must be a useful number of qualified transactions to infer from. 2. The circumstances surrounded each transaction should be known.

PROPERTY BAROMETER FNB Mining Towns House Price Indices

House Price Indexes: Why Measurement Matters

Housing price indexes in Central and Eastern Europe. A comparative study on the models.

PUBLICATION 1905 A Reprint from Tierra Grande

Residential Property Index Series. January 2018

WORKING PAPER NO /R MEASURING HOUSING SERVICES INFLATION. Theodore M. Crone Leonard I. Nakamura Richard Voith

IREDELL COUNTY 2015 APPRAISAL MANUAL

Working Papers. Research Department WORKING PAPER NO. 99-9/R MEASURING HOUSING SERVICES INFLATION. Theodore M. Crone Leonard I. Nakamura Richard Voith

Land Value Estimates and Forecasts for Reston. Prepared for Reston Community Center April 2013

Agricultural FINANCE Monitor

REALTORS GUIDE MLS HOME PRICE INDEX (MLS HPI)

1 February FNB House Price Index - Real and Nominal Growth

Price Indexes for Multi-Dwelling Properties in Sweden

86M 4.2% Executive Summary. Valuation Whitepaper. The purposes of this paper are threefold: At a Glance. Median absolute prediction error (MdAPE)

THE ACCURACY OF COMMERCIAL PROPERTY VALUATIONS

Search for Comparables Workshop on Transfer Pricing ICAI, Bangalore Chapter, 19 August 2006

Direct Capital Value Comparison (Sales Comparison Approach)

3 November rd QUARTER FNB SEGMENT HOUSE PRICE REVIEW. Affordability of housing

Chapter 13. The Market Approach to Value

NBER WORKING PAPER SERIES PRICES OF SINGLE FAMILY HOMES SINCE 1970: NEW INDEXES FOR FOUR CITIES. Karl E. Case. Robert J. Shiller

DATA APPENDIX. 1. Census Variables

Department of Economics Working Paper Series

The TAUREAN Residential Valuation System An Overview

Spatial Dependence in a Hedonic Real Estate Model: Evidence from Jamaica

5. PROPERTY VALUES. In this section, we focus on the economic impact that AMDimpaired

DEPARTMENT OF ECONOMICS WORKING PAPER SERIES. The Demand for Educational Quality: Combining a Median Voter and Hedonic House Price Model

THE EFFECT OF PROXIMITY TO PUBLIC TRANSIT ON PROPERTY VALUES

Glossary of Terms & Definitions

Performance of the Private Rental Market in Northern Ireland

Bargara Property Factsheet

MAAO Sales Ratio Committee 2013 Fall Conference Seminar

The Effect of Relative Size on Housing Values in Durham

Sorting based on amenities and income

General Market Analysis and Highest & Best Use. Learning Objectives

Panos LOLONIS, Greece. Key words: Cadastre, Hellenic Cadastre, Planning, Decision-making, Statistical Estimation SUMMARY

While the United States experienced its larg

Course Residential Modeling Concepts

REDSTONE. Regression Fundamentals.

Following is an example of an income and expense benchmark worksheet:

Monitoring commercial property prices during the crisis

MONTHLY HOUSE PRICE INDEX REPORT

APPLICATION OF GEOGRAPHIC INFORMATION SYSTEM IN PROPERTY VALUATION. University of Nairobi

German experiences with creating real estate databases

What s Next for Commercial Real Estate Leveraging Technology and Local Analytics to Grow Your Commercial Real Estate Business

House prices up by 7.6% on a year before

Comparative Housing Market Analysis: Minnetonka and Surrounding Communities

Washington Department of Revenue Property Tax Division. Valid Sales Study Kitsap County 2015 Sales for 2016 Ratio Year.

Prices of dwellings in housing companies

Prices of dwellings. Prices of dwellings rose in May. 2012, May

METROPOLITAN COUNCIL S FORECASTS METHODOLOGY

Prices of dwellings in housing companies

Prices of dwellings in housing companies

Developing a Residential Property Price Index (RPPI) for Canada: Approach, Risks and Challenges

Transcription:

Marta Widłak / Economic Institute The hedonic house price index for Poland modelling on NBP BaRN data Narodowy Bank Polski International Workshop, Zalesie Górne, 14-15 November 2013

Motivation Unprecedented house price boom (2006-2008) due to the fast mortgage growth showed the necesity of a closer monitoring of the housing market. The new trend in economic macromodelling stresses the importance of microfoundations. New task of central banks macroprudential policy in the aftermath of the financial crises. Lack of reliable official statistics for the housing market. Untill today no official House Price Index (HPI) exists for Poland (not even based on the simple average). Lack of appropriate, reliable data to construct HPI. => NBP decides to conduct its own basic survey of the housing market and compile hedonic HPI. 2

Motivation 3 House prices in Poland in 16 cities and mortgage loans

I focus on: Which hedonic method to choose to calculate HPI for Poland? What practical problems occur during hedonic modelling and how to solve them? Short information of the NBP s survey and BaRN database Hedonic index Warsaw as a special case to analyse different hedonic HPI approaches HPI for 16 voivodioship cities Conclusions 4

5 BaRN database The survey started in the 3q 2006 and is conducted in 16 capital cities of Polish voivodeships. Source: real estate agencies, developers, notary acts (municipall offices) Asking and transaction data on dwellings in multiunit buildings, new construction and existing stock. Since 2011 also the segments of single family houses and plots are monitored. Frequency: quarterly 17 main characteristics + 2 assesments of location + address (without the building number) + 28 proxies for location variables added using GIS (6 biggest cities; accurate for mid points of the streets) Data is gathered by local market analysts from the NBP Branches. They have gained knowledge and experience in their local housing markets, that enables them to correct the data and select true information. They also correct for repetitions of records and very often complete the data. Thus our dataset is probably most complex in terms of housing atributes information as for the whole country. The week point is lack of randomized sample and lack of exact address of the real estate.

BaRN database Number of transaction data in BaRN - higher since the survey became mandatory in 2013. The sample covers around 15-20% of total population. Due to the lack of correct official statistics on housing transactions samples are not random.

Hedonic quality adjustment Quality adjustment seems to be one of the most important issue to address when measuring true price dynamics Mean price, hedonic time - dummy and characteristics structure indices ( Warsaw secondary housing market, transactions, Q/Q)

Hedonic quality adjustment The hedonic index is any index that makes use of a hedonic function (Triplett 2006). There are 6 different methods to compute a hedonic index (ILO 2004). In my work I consider 4 direct methods (which do not require match - models)

Hedonic quality adjustment 9 Imputation index One hedonic model for the reference period used to calculate shadow prices in all periods and then we compare shadow with real prices approach applied by NBP since 2009, separate models for 16 voivodeship cities or separate hedonic models for every period to compare shadow prices relevant for period t+1 with real prices from period t p real house price, z vector of characteristics, dashed p shadow price

Hedonic quality adjustment 10 Characteristic price index Laspeyres type Pasche type q weight for j characteristic; a hedonic coefficient for j characteristic BUT (1) experience reveals lack of statistical significance of coefficients (characteristics prices) in each period hedonic model; (2) the same model specification is not appropriate for every period. How to address these issues?

Hedonic quality adjustment 11 Time dummy index D time - dummies for T periods (pooled approach) or for two adjacent periods (adjacent aproach) Index = exp b t + 0,5σ 2 Critisized for the assumption of stable price characteristics over all pooled periods (demand and supply for housing characteristics remain unchanged which is not consistent with the theoretical model of Rosen 1974).

Warsaw which hedonic approach? 12 Hedonic model for Warsaw 6 structural characteristics + dummies for district + 14 location variables Two step of outliers elimination firstly expert assesment (area, total price, price per sq. m.); secondly according to studentized residuals. Log (price sq.m.) = f(area, sq_area, rooms, finishing standard, ownership law, construction year, district, center dist., boundry dist., metro dist., tram noise, industrial noise, cemetery, discounter, medical centers, hospitals, No green areas, green dist., protected green points dist., industrial points, trade centers dist.)

Warsaw which hedonic approach? 14 HPI for Warsaw (Q/Q)

Warsaw which hedonic approach? 15 HPI for Warsaw (Q/Q)

Warsaw which hedonic approach? 16 HPI for Warsaw (1q 2006 = 100)

Warsaw which hedonic approach? 17 Hedonic HPIs for Warsaw (3 q 2006 = 100)

Warsaw which hedonic approach? 18 Pooled time dummy index for Warsaw (1 q 2009 = 100) No much need for revisions of the index when adding new quarter data

Characteristics price models - Warsaw In case of models for one quarter data (characteristic price approach) most of coefficients are statistically insignificant P-values of hedonic models for characteristic price index. In red are coefficients that are statistically insignificant on 10% level. Etykiety wierszy 20061 20062 20063 20064 20071 20072 20073 20074 20081 20082 20083 20084 20091 20092 20093 20094 20101 20102 20103 20104 20111 20112 20113 20114 20121 20122 20123 20124 20131 20132 20133 l_granica_~t 0,01 0,30 0,21 0,89 0,03 0,06 0,45 0,13 0,11 0,48 0,05 0,11 0,97 0,03 0,73 0,20 0,00 0,03 0,03 0,23 0,04 0,14 0,00 0,06 0,03 0,93 0,25 0,14 0,51 0,83 0,02 l_stacja_m~t 0,32 0,03 0,06 0,17 0,13 0,15 0,08 0,54 0,90 0,81 0,00 0,02 0,11 0,92 0,09 0,46 0,22 0,33 0,50 0,12 0,04 0,03 0,02 0,25 0,17 0,00 0,01 0,00 0,34 0,80 0,02 l_centra_h~t 0,08 0,48 0,60 0,57 0,05 0,84 0,77 0,18 0,01 0,74 0,31 0,07 0,31 0,06 0,10 0,58 0,67 0,90 0,07 0,91 0,89 0,50 0,06 0,36 0,50 0,26 0,64 0,81 0,66 0,50 0,08 halas_tram~y 0,28 0,39 0,58 0,55 0,08 0,10 0,00 0,86 0,70 0,21 0,09 0,22 0,94 0,83 0,47 0,43 0,68 0,04 0,47 0,32 0,02 0,20 0,83 0,44 0,12 0,06 0,93 0,22 0,23 0,34 0,31 przemysl_h~s 0,82 0,00 0,36 0,21 0,74 0,68 0,64 0,68 0,26 0,55 0,14 0,00 0,20 0,93 0,80 0,33 0,74 0,02 0,01 0,75 0,12 0,39 0,00 0,44 0,17 0,11 0,12 0,88 0,05 0,78 0,10 zielen_l1000 0,25 0,00 0,34 0,09 0,14 0,20 0,84 0,20 0,92 0,06 0,09 0,00 0,00 0,59 0,44 0,45 0,82 0,00 0,51 0,55 0,65 0,21 0,23 0,32 0,14 0,24 0,28 0,92 0,56 0,04 0,16 zdrowie~1000 0,00 0,30 0,81 0,13 0,80 0,41 0,26 0,01 0,17 0,11 0,39 0,02 0,49 0,77 0,85 0,13 0,01 0,75 0,85 0,99 0,00 0,29 0,00 0,09 0,37 0,05 0,00 0,43 0,34 0,30 0,06 powierzchnia 0,57 0,87 0,03 0,83 0,54 0,00 0,28 0,78 0,16 0,92 0,13 0,07 0,22 0,98 0,59 0,97 0,09 0,06 0,05 0,68 0,00 0,00 0,01 0,14 0,00 0,00 0,21 0,00 0,10 0,33 0,00 sq_pow 0,18 0,29 0,12 0,47 0,89 0,00 0,34 0,80 0,05 0,83 0,22 0,05 0,35 0,95 0,28 0,97 0,11 0,06 0,40 0,78 0,01 0,00 0,00 0,52 0,00 0,00 0,69 0,00 0,16 0,79 0,00 ilepokoi_1_2 0,36 0,26 0,59 0,75 0,91 0,60 0,79 0,71 0,26 0,84 0,30 0,52 0,44 0,84 0,29 0,94 0,10 0,28 0,98 0,12 0,12 0,28 0,29 0,49 0,31 0,63 0,01 0,13 0,59 0,43 0,55 stand_1 0,66 0,12 0,00 0,22 0,36 0,02 0,61 0,40 0,00 0,03 0,01 0,03 0,13 0,38 0,93 0,07 0,16 0,16 0,13 0,07 0,16 0,29 0,00 0,00 0,01 0,01 0,62 0,00 0,01 0,00 0,00 stand_3_4 0,36 0,52 0,45 0,16 0,21 0,56 0,01 0,00 0,03 0,78 0,73 0,89 0,03 0,09 0,76 0,65 0,56 0,10 0,16 0,10 0,01 0,00 0,26 0,01 0,12 0,03 0,00 0,00 0,75 0,14 0,70 spoldzielcze 0,08 0,27 0,48 0,30 0,85 0,37 0,21 0,34 0,03 0,69 0,45 0,47 0,11 0,69 0,04 0,74 0,02 0,00 0,82 0,63 0,24 0,71 0,81 0,31 0,34 0,68 0,19 0,92 0,93 0,00 0,09 rok_1_2 0,99 0,02 0,96 0,88 0,54 0,73 0,81 0,00 0,73 0,50 0,96 0,45 0,23 0,14 0,36 0,25 0,02 0,72 0,24 0,84 0,74 0,98 0,90 0,00 0,10 0,75 0,71 0,52 0,70 0,08 0,30 rok_3 0,18 0,03 0,08 0,04 0,72 0,17 0,16 0,49 0,23 0,28 0,00 0,09 0,01 0,03 0,00 0,32 0,02 0,26 0,02 0,44 0,01 0,00 0,00 0,00 0,00 0,05 0,01 0,06 0,03 0,64 0,00 rok_4 0,01 0,00 0,00 0,20 0,05 0,12 0,05 0,23 0,20 0,02 0,00 0,00 0,01 0,00 0,01 0,09 0,01 0,02 0,00 0,33 0,00 0,00 0,00 0,00 0,14 0,00 0,05 0,02 0,09 0,15 0,00 rok_5 0,04 0,00 0,00 0,08 0,01 0,07 0,04 0,84 0,52 0,05 0,00 0,00 0,40 0,00 0,01 0,12 0,00 0,00 0,01 0,05 0,01 0,02 0,00 0,00 0,11 0,00 0,37 0,10 0,05 0,00 0,00 rok_7 0,53 0,49 0,77 0,15 0,00 0,81 0,02 0,09 0,95 0,04 0,06 0,68 0,37 0,63 0,56 0,21 0,85 0,35 0,22 0,27 0,73 0,04 0,30 0,34 0,88 0,07 0,01 0,01 0,76 0,01 0,00 rok_8 ###### ###### 0,53 0,07 0,94 0,91 0,18 0,00 0,22 0,00 0,46 0,73 0,68 0,55 0,09 0,54 0,19 0,33 0,34 0,00 0,20 0,01 0,01 0,22 0,16 0,01 0,00 0,01 0,15 0,00 0,02 rok_nowe 0,29 0,40 0,36 1,00 ###### ###### ###### ###### 0,25 0,15 0,42 0,36 0,95 0,67 0,12 0,71 0,66 0,01 0,00 0,33 ###### 0,45 0,81 0,91 ###### 0,02 0,95 0,16 ###### ###### 0,01 gr1 0,00 0,01 0,00 0,00 0,17 0,21 0,01 0,04 0,09 0,01 0,06 0,14 0,07 0,09 0,00 0,91 0,98 0,72 0,00 0,00 0,09 0,00 0,05 0,00 0,09 0,00 0,13 0,00 0,27 0,02 0,00 gr2 0,09 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,01 0,00 0,01 0,24 0,05 0,01 0,00 0,00 0,01 0,02 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,10 0,00 0,00 gr3 0,04 0,02 0,00 0,00 0,02 0,02 0,01 0,07 0,01 0,05 0,08 0,49 0,10 0,00 0,08 0,69 0,08 0,30 0,00 0,00 0,00 0,01 0,00 0,00 0,00 0,00 0,06 0,03 0,41 0,00 0,00 gr4 ###### 0,00 ###### 0,00 ###### 0,00 ###### ###### 0,30 0,00 ###### ###### ###### ###### 0,00 0,00 0,15 0,01 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 ###### 0,00 0,00 gr5 0,19 0,11 0,00 0,14 0,29 0,60 0,00 ###### 0,66 0,00 0,02 0,06 0,59 0,00 0,00 0,93 0,31 0,02 0,10 0,22 0,01 0,01 0,19 0,01 0,08 0,01 0,37 0,02 0,55 0,00 0,00 mokotow 0,01 0,32 0,04 0,01 0,00 0,02 0,47 0,46 0,20 0,04 0,01 0,11 0,83 0,02 0,03 0,19 0,88 0,59 0,00 0,10 0,09 0,44 0,19 0,01 0,18 0,39 0,66 0,00 0,92 0,05 0,00 ochota 0,76 0,80 0,06 0,03 0,02 0,07 0,56 0,58 0,76 0,09 ###### 0,13 0,46 ###### 0,27 0,75 0,97 0,67 0,10 0,32 0,32 0,07 0,26 0,04 0,01 0,36 0,67 0,23 0,65 0,57 0,00 wola 0,97 0,57 0,01 0,12 0,05 0,26 0,42 0,08 0,08 0,28 0,10 0,29 0,57 0,40 0,42 0,66 0,15 0,11 0,00 0,03 0,01 0,09 0,04 0,01 0,01 0,03 0,05 0,00 0,28 0,01 0,00 ursynow 0,26 0,31 0,05 0,00 0,00 0,05 0,00 0,08 0,82 0,14 0,02 0,11 0,06 0,10 0,05 0,52 0,95 0,51 0,06 0,00 0,07 0,02 0,18 0,02 0,01 0,00 0,29 0,00 0,49 0,46 0,00 zoliborz 0,11 0,60 0,00 0,00 0,99 0,02 0,67 0,19 0,02 0,89 0,16 0,88 0,29 0,27 0,04 0,72 0,85 0,80 0,00 0,04 0,22 0,35 0,63 0,69 0,00 0,00 0,83 0,08 0,73 0,00 0,01 wilanow 0,00 0,74 ###### ###### ###### 0,68 ###### ###### ###### 0,03 ###### 0,18 0,03 ###### 0,00 0,81 0,00 0,05 ###### ###### 0,09 0,05 0,00 0,04 0,00 0,09 0,15 0,03 ###### 0,05 0,00 _cons 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00

Characteristics price models - Implicit prices - Warsaw 20 and changes substantially from period to period and have wrong signs although the RESET test proves a good specification for most of the models.

Characteristics price models - Implicit prices - Warsaw 21

Characteristics price models - Implicit prices - Warsaw 22

Characteristics price models - Implicit prices - Warsaw 23

Characteristics price models - Implicit prices - Warsaw 24

Time dummy hedonic HPI for 16 main cities 25 Hedonic model for 6 and 10 cities 2 separate models for 6 bigest cities and the group of 10 cities (small differences in specifications) 9 structural characteristics + dummies for cities Two step of outliers elimination firstly expert assesment (area, total price, price per sq. m.); secondly according to studentized residuals. Log (price sq.m.) = f(no of rooms (6 cities) or area (10 cities), sq_area (10 cities), location assessment, finishing standard, building technology, floor, No of floors, kitchen type, ownership law, construction year, dummy for each city) Problem with specification test Ramsey s RESET but all structural variables significant at 1% level. [No of obs. - 10 583 for 6 cities and 14 861 for 10 cities]

Pooled time dummy index for Warsaw (Q/Q) 27

Pooled time dummy index for 16 cities (3 q 2006 = 100) 28

Conclusions and questions for discussion Need to deal with data selection bias (randomization of samples needed but unknown population of transactions) Hedonic approach gives more reliable index than simple average or median but which hedonic approach to apply? Detailed analysis of the Warsaw market shows that imputation and pooled time dummy indices are preferable. They give similar results despite differences in assumptions (!) Why? What s your experience? Why most preferable characteristic price approach is not valid? Can we suspect any systematic error in our econometric job? Pooled time-dummy approach chosen for compsite index due to operational reasons (no need for re-specification and reestimation of models on quarterly basis). I am not aware of stability of implicit price coeff. but should the structure of characteristics should be somehow controlled in this case?

Conclusions and questions for discussion Big hedonic models (pooled TD Warsaw, 6 cities, 10 cities) sgnificant coeff., reasonable values and good signs but problem with specification tests (Ramsey s RESET or linktest). Small hedonic models (characteristic price approach Warsaw) in reverse Why? How specify such models? Are the choosen specification tests appropriate? Exmining more sophistocated econometric modeling tools (GAM approach and spatial modelling) Work on statistical formula of the index (Fisher type? weighted estimator for time-dummy method?) HPI for primary market and single family houses (one composite HPI for all segments) Other suggestions from the floor most welcome! 30