Hierarchical Trend Model
Marc K. Franckemfrancke@ortax.nl
1Department of Econometrics
Vrije Universiteit Amsterdam
2OrtaX
Amsterdam
Advances in mass appraisal methods
Outline
1 State-Space Model
2 The Hierarchical Trend Model
State-Space Model
The linear Gaussian State-Space Model:
yt = Ztαt+ εt, εt ∼ N(0, σ2Ht), αt+1 = Ttαt+Rtηt, ηt ∼ N(0, σ2Qt), for t=1, . . . ,T , initialized by α1 = a0+A0β +R0η0, η0 ∼ N(0, σ2Q0), β ∼ N(β0, σ2Σ), yt is an(nt×1)vector of observations
αt is an unobserved state vector
εt, ηt,andβ are uncorrelated
The matrices Zt, Tt, Ht,Rt and Qt are called system
matrices; they may depend on unknown parameters
Examples of State-Space Models
Simple State-Space Modelyt = Ztαt + εt, εt ∼ N(0, σ2I), αt+1 = αt + ηt, ηt ∼ N(0, σ2ηI),
Forση =0, this is a regression model yt =Ztα + εt
Forση >0, time-varying coefficients
When Zt =1, then for
ση= ∞, a dummy variable model
ση=0,αt is constant in time
State-Space Model and the Kalman filter
State-Space models can efficiently be estimated by the Kalman filter
Estimates of the state-vector, conditional on the parameters in the system matrices
The likelihood function for the parameters in the system matrices
Extensions
Non-Gaussian disturbances, for example t-distribution Nonlinear relations
yt = Zt(αt) + εt, αt+1 = Tt(αt) +Rtηt,
Outline
1 State-Space Model
2 The Hierarchical Trend Model
Market Segments
Dependent variable yt : ln(Selling Prices) at time t
A priori classification ofmarket segments(MS):
Clusters ofneighbourhoods(districts) Clusters ofhouse types
Example for the Heerlen case
House type Description Cluster of House Type
1111 Detached house 1
1113 Detached bungalow 1
1115 Detached converted farmhouse 1
1121 Semi-detached house 1
1123 Semi-detached bungalow 1
1125 Semi-detached converted farmhouse 1 1128 Semid-detached drive-in house 1
1131 Row house 2 1133 Row bungalow 2 1138 Row drive-in 2 1141 Corner house 2 1143 Corner bungalow 2 1171 Linked house 1 1173 Linked bungalow 1
1175 Linked converted farmhouse 1
Price development
In the same MS houses have the same price development
The trend for MSjk is the sum of
Generaltrend:µt
Districttrend:ϑj,t (in deviation from the general trend)
House type clustertrend:λk,t (” ”)
Houses in the same MS have the same price development (µt+ ϑj,t+ λk,t)
Time index t is in months
Time Trends
General Trend: local linear trend model
µt+1= µt + κt + ηt, ηt ∼N(0, σ2µI) κt+1= κt + ζt, ζt ∼N(0, σ2κI)
Cluster Trends are modeled as random walks
λt+1= λt + ςt, ςt ∼N(0, σ2λI) ϑt+1= ϑt + ωt, ωt ∼N(0, σϑ2I)
σ2η σ2ζ Model
∞ Dummy variable model 0 0 Deterministic linear trend
0 Random walk with drift >0 > 0 Level and slope may vary over time
Outline
1 State-Space Model
2 The Hierarchical Trend Model
Explanatory variables (X
t)
Linear specification makes no sense:
y = β1ln x1+ β2ln x2+ α + ,
Y =xβ1
1 ×x
β2
2 ×exp(α + ),
Internal floor space: x1; Lot size: x2
Multiplication of internal floor space and lot size
Nonlinear specification, like for example
y =ln(xβ1 1 exp[z 0δ] + β 2x2) + α + , Y = (xβ1 1 exp[z 0 δ] + β2x2) ×exp(α + ), where z contains variables regarding the internal floor
space x1, like year of building and maintenance
Internal floor space and Lot size
Forβ1<1, Y is less than proportional to x1
Lot size
Discontinuities can be circumvented by the use of an exponential function
Corrected lot size
Lot size and neighbourhood effects
Correction function for lot depends on
location(city center or rural area)
house type(detached house or row house)
Neighbourhoods
Neighbourhoods in the same district have thesame trend
but can havedifferent levels
Outline
1 State-Space Model
2 The Hierarchical Trend Model
Market Segments Explanatory variables Model Specification 3 Results Data Estimates Extensions
The Hierarchical Trend Model
yt =iµt +Dϑ,tϑt +Dλ,tλt+Dφ,tφ +f(Xt, β) + t, t ∼N(0, σ2I) µt+1 = µt+ κt + ηt, ηt ∼ N(0, σµ2), κt+1 = κt + ζt, ζt ∼ N(0, σκ2), ϑt+1 = ϑt + ωt, ωt ∼ N(0, σϑ2I), λt+1 = λt + ςt, ςt ∼ N(0, σλ2I), φ ∼ N(0, σφ2I)the scalarµt thegeneraltrend
the vectorϑt thedistricttrend
the vectorλt thehouse type clustertrend
Outline
1 State-Space Model
2 The Hierarchical Trend Model
Market Segments Explanatory variables Model Specification 3 Results Data Estimates Extensions
Data Heerlen
2658 observations from January ’01 until December ’04 52 neighbourhoods: 6 districts
15 house types: 2 house type clusters
Outline
1 State-Space Model
2 The Hierarchical Trend Model
Market Segments Explanatory variables Model Specification 3 Results Data Estimates Extensions
Model results
Variable Coef T-value
House size in m3 0.7339 49.02 Age -0.0075 -22.69 Year of construction : 1920 - 1945 -0.3980 -25.85 Year of construction : 1900 - 1920 -0.3971 -13.11 Year of construction < 1900 -0.2559 -4.34 Poor Maintenance -0.3112 -5.20 Good Maintenance 0.0795 3.45 Lot size in m2 0.1230 10.85 Number of ’dormers’ 3.7688 3.88 Garage in m3 0.1501 7.87 Carport in m2 0.2975 3.12 Sun room in m2 0.1315 2.50 Cellar in m3 0.0913 4.79 Detached house 0.1074 8.90 Linked house 0.0597 4.53 Corner house 0.0292 4.18 Bungalow 0.1354 8.08
An increase of house size with 10%, gives an increase of value with 7.3%
Difference between ’poor’ and ’good’ maintenance is 0.4, approximately 49% of the house value excluding the land
Model results
Fit of model measured byσ, can be interpreted as a
standard deviation (sd) of 12%
The sd of the general trendσµis monthly 0.629%,
corresponding with 2.2% a year
The sd of the district trendσϑis monthly 0.401%,
corresponding with 1.4% a year
The sd of the house type cluster trendσλ is negligible
The sd of the cluster trendσφis 10.7%
Variable Coef σ 0.1200 σµ 0.00629 σκ 0.00039 σϑ 0.00401 σλ 0.00032 σφ 0.10695 LL 1876.3501
Outline
1 State-Space Model
2 The Hierarchical Trend Model
Market Segments Explanatory variables Model Specification 3 Results Data Estimates Extensions
Model Extensions
Heteroskedasticity (over for example house type) Regression parameters varying over
time space
Spatial correlation in district trends
Summary
Hierarchical Trend Model (HTM) is aState-Space model.
State-Space model is aparametricmodel.
State-Space model allows forflexible functional forms.
(stochastic trends)
The HTM can be characterized as as
Hedonic PriceModel
Time SeriesModel
SpatialModel
Hierarchical Model orMultilevelmodel