Know What You Are Looking For

1
Know What You Are Looking For

• A Theoretical Framework for Hedonic Office Studies

Philip Koppels, Hilde Remøy and Hans de Jonge
2
Introduction
Accommodation Preferences of Office Users

• Buyers market conditions what do office users
  prefer?
• Hedonic office rent studies since the 1980s
• Theoretical underpinning interpretation of
  variables
• Statistical issues multicollinearity, unequal
  variance, .
• Structure of the presentation
• Research introduction
• Modeling approach
• Hedonic analysis
• Interpretation of results

3
(No Transcript)
4
(No Transcript)
5
Modelling Approach
A Linear Mixed Model

• Number of observations Thin markets
• Longer time series or larger research area
• Multiple correlated transactions in one building
  repeated measurements?
• Two level hierarchical model
• Level one transaction
• Level two the building
• Other issues
• Unbalanced panel data, unequal spaced
  observations
• Spatial autocorrelation

6
Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics
Assumption Coefficient Observed Expected Std Dev Z Pr gt Z
Randomization Moran's I -0.129 -0.00144 0.0189 -6.74 lt.0001
Randomization Geary's c 1.360 1.00000 0.0862 4.17 lt.0001
7
Basic Location Adjusted Model
Solution for Fixed Effects Solution for Fixed Effects Solution for Fixed Effects Solution for Fixed Effects Solution for Fixed Effects Solution for Fixed Effects
Effect Estimate Standard Error DF t Value Pr gt t
Intercept -2.3545 1.1228 20 -2.10 0.0489
CPI 96 1.5736 0.2242 15.5 7.02 lt.0001
VACANCY RATE (LAG2) -0.02133 0.004055 9.45 -5.26 0.0004
AGE -0.00882 0.003132 175 -2.82 0.0054
LOG DISTANCE TO IC -0.05605 0.02993 148 -1.87 0.0631
LOG TRAVEL TIME HIGHWAY -0.02480 0.04784 150 -0.52 0.6049
LOG EMPLOYMENT FB 0.05961 0.02829 159 2.11 0.0367
LOG EMPLOYMENT INDUSTRY -0.00007 0.000023 151 -3.28 0.0013
LOG FACILITIES 0.07427 0.01722 151 4.31 lt.0001
SITE QUALITY 0.1347 0.03806 141 3.54 0.0005

• Model fit statistics
• BIC first model -659.4
• BIC second model -784.7

8
Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics Autocorrelation Statistics
Assumption Coefficient Observed Expected Std Dev Z Pr gt Z
Randomization Moran's I -0.127 -0.00147 0.0190 -6.60 lt.0001
Randomization Geary's c 1.470 1.00000 0.0852 5.52 lt.0001
9
Final Results
Covariance Parameters

• Random building intercept significant
• Random age coefficient significant
• Covariance building intercept and age
  coefficient negative
• Range spatial correlation 550 meters

10
Final Results
Fixed Effects
Effect Estimate Std. Error Pr gt t
Intercept -1.9051 1.1014 0.1005
CPI 96 1.5084 0.223 lt.0001
VACANCY RATE (LAG2) -0.02185 0.004067 0.0004
AGE -0.00761 0.003037 0.0132
LOG DISTANCE TO IC -0.03232 0.02822 0.2541
LOG TRAVEL TIME HIGHWAY -0.02319 0.04398 0.599
PARKING LOTS 0.03391 0.02265 0.1368
LOG EMPLOYMENT FB 0.04154 0.02495 0.0982
LOG EMPLOYMENT INDUSTRY -0.00009 0.000021 lt.0001
LOG FACILITIES 0.06291 0.01604 0.0001
SITE QUALITY 0.07851 0.03486 0.0262
11
Final Results
Fixed Effects
Effect Estimate Std. Error Pr gt t
PARKING FACILITIES 0.1128 0.03603 0.0022
FAÇADE BRICKS 0.0541 0.03785 0.1552
FAÇADE NATURAL STONE 0.1445 0.04877 0.0037
FAÇADE GLASS 0.1143 0.04974 0.0232
COMPANY LOGO 0.06452 0.02979 0.0322
LOG RECEPTION AREA GFA 0.03293 0.01761 0.0638
ELEVATOR RATIO 201.03 85.0416 0.0195
LAY-OUT FLEXIBILITY LOW 0.01944 0.03286 0.5552
DAY LIGHT lt50 FAÇADE -0.07796 0.03571 0.0309
FLOOR HEIGHT LOW -0.07758 0.09198 0.4003
12
Questions? Contact author p.w.koppels_at_tudelft.nl