Exploring the Case for Expert-Defined Housing Submarket Boundaries

1

• Exploring the Case for Expert-Defined Housing
  Submarket Boundaries
• Berna Keskin and Craig Watkins
• University of Sheffield

2
Introduction Aim Objectives

• Aim
• to explore the merits of expert-defined
  submarket boundaries when compared with
  submarkets constructed using other statistical
  methods.
• Objectives
• how should analysts seek to construct submarkets
  if they are operating in a market where the
  quality and availability of housing transactions
  datasets is limited?
• Approach
• The use of prior knowledge Principal components
  analysis (PCA) combined with cluster analysis and
  definitions based on the views of (expert) real
  estate professionals. The performance of these
  approaches is compared in terms of their impact
  on the accuracy of hedonic price estimates.
• measuring the impact on standard error
• computing the predictive accuracy

3
Motivation of the Study

• Segmented Market structure
• Housing market in Istanbul are highly segmented
• There are significant price differences, in
  different parts of the market for homes with the
  same physical features and locational attributes
• Population 10,033,478.
• Istanbul population/Turkey 14.78 in 2000
  (TUIK,2006), surpasses the population of 22 EU
  countries (Eurostat).
• 2,550,000 households and 3,391,752 housing units
• The problems
• high increase rate in population,
• the gap in the incomes
• lack of enough amounts of residential plots.
• land rent and speculation.

4
Submarket Delineation (A priori- PCACluster
Analysis-Experts)

• A priori segmentations which are considered to
  be the most probable. (5 submarket)
• Principal components analysis (PCA) combined
  with (K means) cluster analysis, (5 submarket)
• Consultation with real estate agents and valuers
  working in the Istanbul market.
• eight semi-structured interviews conducted in
  November 2007.
• spatial submarket boundaries on a 1/200,000 scale
  map
• the interviewees drew between five and seven
  submarkets, even though no guidance was provided
  and no restrictions were set.

5
An example of the experts submarket
identification
6
The synthesis map of experts map
7
The synthesis map of experts map
8
Data
Variables
Property Characteristics
Socio-economic Characteristics
Neighbourhood Characteristics
Locational Characteristics
1.Housing Type 2. Rooms 3. Floor Area 4. Elevator
5. Garden 6. Balcony 7. Storey 8. Site 9. Age

• Income
• Household size
• Living period in the neighbourhood
• Living period in Istanbul

• Satisfaction from
• School
• Health service
• Cultural facilities
• Playground
• Neighbour
• Neighbourhood quality

• Earthquake risk
• Continent
• Travel time to shopping centres
• Travel time to jobs and schools

Italic variables are excluded due to
multicollinearity.
9
Comparison of Models

10
Comparison of Models
Basic Hedonic Model P f ( Fa, I, Lp, -Eq, S,
A, Ls,) Fa Floor Area S Site A
Age Ls Low Storey I Income of the
household Lp Living Period in Istanbul Eq
(-)Earthquake Damage Rsquare 0.60
Hedonic Model with a priori Submarket
Variables P f ( Fa, I, Lp, -Eq, S, A, C, N,Sm1,
Sm3, -Sm4, -Sm5) Fa Floor Area S Site A
Age C Continent I Income of the
household Lp Living Period in Istanbul N
Neighbor satisfaction Eq (-)Earthquake
Damage Sm1 1st submarket Sm3 3rd
submarket Sm4 (-)4th submarket Sm5 (-)5th
submarket Rsquare 0.67
Hedonic Model (experts) submarket variables P
f ( Fa, Ls, Lp, HS, Sm1, -Sm3, -Sm4, -Sm5 Fa
Floor Area S Site Lp Living Period in
Istanbul Hs Household size Sm1 1st submarket
Sm3 (-)3rd submarket Sm4 (-)4th
submarket Sm5 (-)5th submarket Rsquare 0.68
Hedonic Model with Cluster Submarket (PCA)
Variables P f (Ls,I, Lp,N,Sm2, Sm3, Sm4,) Ls
Low Storey I Income of the household Lp
Living Period in Istanbul N Neighbor
satisfaction S Site Sm2 2nd submarket Sm3
3rd submarket Sm4 4th submarket Rsquare
0.61
11
RMSE Test

12
Accuracy Test

13
Conclusions
The specification based on prior knowledge led to
the greatest reduction in standard error (at more
than 20). The expert-defined formulation reduced
the standard error by just over 15 The
predictive accuracy test showed that the
expert-defined submarket formulation produced the
largest proportionate decrease. It also generated
the largest proportion of estimates within ten
and twenty per cent of the actual value with more
than 20 and 40 in the respective bands. The
results do not provide comprehensive evidence
that expert-defined submarkets are superior to
specifications based on alternative methods.
The expert-defined model does, however, perform
well in terms of predictive the submarkets
constructed are a reasonable approximation of the
true submarket structure. These findings
suggest that the methods used here to consult
expert and construct a consensus view might offer
a reasonable solution to analysts operating in
markets where data availability is limited.