Market value of real estate

DOCUMENTATION PATRIMONIALE

PATRIMONIUMDOCUMENTATIE

1

Market value of real estate Guy Hofman Gérald Henskens

October 9th, 2014 Federale Overheidsdienst FINANCIEN Service Public Fédéral FINANCES



2

Outline • • • • • •

Context and issue Solution Increasing performance Watch out Return on investment Future improvements

Federale Overheidsdienst FINANCIEN Service Public Fédéral FINANCES



3

Contexte and issue • One of the missions of Patrimony Documentation = fiscal control on declared value of real property in deeds and declarations


To prevent and fight fraud in the matter of registration duties and inheritance taxes



4

Contexte and issue • Detection of dubious property transactions • Establish correct market value of conveyed property  different categories : land, house, apartment, …  situation : anywhere in Belgium  indication of priority for control




5

Solution • Modelling real property data  only houses and apartments  multiple regression with 11 objective parameters o o o o o

Situation : street or zone Type : villa, bel-étage, serial construction, bungalow, … # facades # floors Year of construction




6

Solution • Modelling real property data  multiple regression with 11 objective parameters o o o o o o

Quality of construction # garages # entities for families # rooms (apartment) or Surface area (house) Total useable floor surface Date of valuation




7

Solution • Elimination of extremes • 1 formula for every municipality blok-zonegeo-zoneval

afdeling

nsr

cc1

cc2

cc3

cc4

cc5

cc7

cc8

cc11

cc13

cc14

perc.opp.

datum

405804

11001

310

40

a

2

n

1935

n

0

3

225

371

546

23/04/2014

00000405804 00405804 11001H00405804 279.715

11001 285.225

VERKOOPWAARDE Vork Laag

323.648 AARTSELAAR

AARTSELAAR AARTSELAAR

0 02

11001

11001

AARTSELAAR AARTSELAAR 1100100310 AARTSELAAR 1100100310 11001 280.000

AARTSELAAR 40

AARTSELAAR a

AARTSELAAR

AARTSELAAR

283.561

280.000

AARTSELAAR 40

AARTSELAAR A

283.561

279.715

260.394

247.192

SEE

intercept

leeftijdsq

cc14_sum

23440,92213

171426,282 1

12,53121762

1285,136009

2n

AARTSELAAR n

AARTSELAAR 0

AARTSELAAR 3

2N

1.935

AARTSELAAR N

AARTSELAAR 0

AARTSELAAR 3

268.450

85

285.225

189.133

340.000

day1

day1sq

leeftijd

prcc1

prcc11

prcc2

0

0

2943,4647 6

0

0,355347881

0,254264912

tuinopp. 321

321 371

546

prcc34

prcc7

prcc8

0,347336 94

0

0

4496

137.641

298.116

7.225

20.214.0 16

Coeff.

gem

AARTSELAAR


cc14sq oppAkte3

-1,710614027

92,462411 83

oppakte3s prgembl q okzone

0

0,30745 12

323.648

Vork Hoog

297.050

350.247

0,92

1,08



8


Solution



9

Solution • Calculated market value with minimum/maximum based on Standard Error of the Estimate • Every 2 weeks import of new transaction data • Selection of possibly dubious transactions with theoretical return




10

Increasing performance • SAS environment with training and coaching • ‘Translation’ of the mathematical model using E-guide • Recruitment of dataminers • Automation • Support of external consultants




11

Watch out • • • • •

Business understanding Data understanding Data quality and cleaning Sponsoring by managers for implementation Feedback from users




12

Return on investment • Potentially recovering 200 million euros in lost property taxes per year (only houses and apartments) Code


Transacties juni 2013 to december 2013 Statuscode betekenis Aantal verkopen Percent verkopen Aantal erfenissen Percent erfenissen 1 Geen controle venale waarde, op basis van type inschrijving 46 0,03% 126 0,15% 2 Geen controle venale waarde, op basis van verstreken termijn 1.061 0,80% 2.355 2,72% 3 Geen controle venale waarde, op basis van type opsteller 2.905 2,20% 515 0,59% 4 Geen controle venale waarde, op basis van voorafgaande schatting 0 0,00% 1.964 2,27% 5 Geen controle venale waarde, op basis van vast recht 544 0,41% 0 0,00% 6 Geen controle venale waarde, op basis van nultarief 4.059 3,07% 4.810 5,55% 41 Te controleren, Geen mathematisch model beschikbaar 48.681 36,81% 32.058 37,00% 47 Te controleren, Mathematische waarde niet berekend 8.989 6,80% 7.498 8,65% 51 Niet te controleren, Locowaarde boven bovenkant vork 18.575 14,04% 4.646 5,36% 52 Niet te controleren, Locowaarde binnen vork 30.003 22,69% 14.464 16,69% 61 Niet te controleren, Locowaarde gelijk onderkant vork 694 0,52% 511 0,59% 62 Te controleren, Locowaarde kleiner dan onderkant vork, geen tarief beschikbaar 137 0,10% 204 0,24% 63 Niet te controleren, Theoretisch rendement kleiner dan 400 euro 2.194 1,66% 3.302 3,81% 64 Te controleren, Theoretisch rendement tussen 400 en 1000 euro 3.103 2,35% 3.477 4,01% 65 Te controleren, Theoretisch rendement groter dan 1000 euro 11.267 8,52% 10.714 12,37% 132.258 100,00% 86.644 100,00%



13

Return on investment • Dissuasive effect : unwanted behaviour decreases dramatically when people realise how we select and control




14

Future improvements • More input from real estate agents and Regions > from 11 to 25 parameters o o o o o o o

Energy Performance Certificate Solar panels Kitchen equipment Sanitary equipment Urbanistic possibilities Mobility possibilities Level of maintenance




15

Future improvements • Extension of the ‘population’ > all buildings • Use of another technique > State Space Modeling > Dutch example ‘Hierarchical Trend Model’




16

Hierarchical Trend Model • Used by Dutch administrations for tax purposes • Based on Time Series • Explains the value of real estate by housing characteristics, location and time • Hierarchy of trends: – Common trend – Location based trend – Building type trend




17

Hierarchical Trend Model • Location based and building type trends modelled as deviations from the common trend • Housing characteristics are time-invariant • Log(price) = influence of characteristics + level common trend + level location based trend + level building type trend + error term


SAS Forum

Twitter Contest – Tweet to win prizes!

6. Het jaarlijks bedrag aan vastgoedfraude in België wordt - alleen voor huizen en appartementen - geschat op 200 miljoen euro B. 20 miljoen euro C. 2 miljard euro Tweet your answer: A.

Example:

Prizes to win:

@spicyanalytics 6C

Start of your tweet

C op yr i g h t © 2 0 1 3 , S A S I n s t i t u t e I n c . A l l r i g h t s r es er v e d .

Question #

Your answer

1st prize: a ticket for Analytics 2015 2nd prize: a book of Prof Bart Baesens: “Analytics in a big data world” 3rd to 30th prize: chocolates with pepper

Winners will be contacted post-Forum !

Market value of real estate

Recommend Documents