PowerPoint - ERES Digital Library

Masaki Mori
Seow Eng Ong
Joseph T. L. Ooi
European Real Estate Society 20th Annual Conference
Vienna, Austria
July 3-6, 2013
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
150
.2
Background
50
Mid Niigata
Prefecture
Earthquake, 2004
0
-.2
-.1
0
above_4
100
.1
Great East Japan
Earthquake, 2011
2004m12005m12006m12007m12008m12009m12010m12011m12012m1
JREIT index return
above_4
2
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
20
0
-.2
5
10
-.1
15
0
above_4
.1
.2
Background (without extremes)
2004m12005m12006m12007m12008m12009m12010m12011m12012m1
JREIT index return
above_4
3
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Research question

Does earthquake risk perception affect return
of J-REIT stocks?

Do investors alter their attention allocation
once earthquake events provoke their fear
and awareness toward earthquake?
4
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Motivation from finance literature
Rare events and asset pricing

A small probability of very bad things
happening affects asset pricing dynamics
(e.g., Rietz, 1988 and Barro, 2005)
Limited attention
Investors show limited attention to relevant
information
 Investors shift attention when faced with
greater uncertainty (e.g., Hirshleifer and Teoh, 2003)

5
Introduction
Data &
Methods
Research
Framework
Results
Conclusions
Why earthquake risk?
 Unpredictable,
Year/
Seismic
intensity
2004
2005
2006
2007
2008
2009
2010
2011
Total
exogenous, and immediate
5+
6-
6+
7 Total
11
5
0
2
0
0
0
17
36
2
2
0
1
1
1
0
4
16
2
0
0
2
1
0
0
4
10
1
0
0
0
0
0
0
1
2
(Japan Meteorological Agency)
16
7
0
5
2
1
0
26
64
6
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Why J-REITs?


Portfolio level earthquake resistance measure
in semi-annual financial reports
(PML: Probable Maximum Loss)
Price information
 We can examine dynamic effects of earthquake
event on prices
7
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Contributions

Implications for limited attention finance
literature
 Hopefully, will shed new light on the impacts of investors’ limited
attention and attention allocation on asset price dynamics
 Implications
for REIT markets
 How earthquake risk is priced in REITs
 REIT markets are being developed in some
earthquake prone countries
8
Research
Framework
Introduction
Data &
Methods
Results
Conclusions
Data
 Study
period from Jan. 2004 to Apr. 2012
 Earthquake
event data from the Japan
Meteorological Agency
 43
REITs
 Price data from DataStream and Association for
Real Estate Securitization in Japan
 PML and financial information from REIT DB
and semi-annual financial reports
9
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Methods

Outlier Robust regression (MM-estimation)
 Dependent variable
○ J-REIT Index total return
 Independent variable
○ # of earthquake equal to or greater than 5 strong in a month (log)
○ Google score as a robustness check

Sub portfolio analysis (1: low 2: middle 3: high)
 PML score
 Financial institution ownership
 Foreign investor ownership
10
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Methods

Event study
 Earthquakes in Japan from Jan. 2004 to April. 2012
○ “LARGE” --- 5-strong (26) and 6-weak (9)
○ “HUGE” --- 6-strong (3) and 7 (2)
Estimation period = -60 to -30 days
 Event window = 0 to +10 days
 Robust test

 Abnormal return serial correlation
 Event-induced volatility
(Kolari & Pynnonen, 2010)
11
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Total return vs. # of quakes
***, **, *, for 1%, 5%, and 10%
significance, respectively
Panel A:
Overall J-REIT index (PML: 5.46)
Constant
# of earthquakes
Estimate z -value
0.015
2.65
-0.019 -2.36 **
Panel B:
Low PML portfolio (safe: 2.96)
Constant
# of earthquakes
Estimate z -value
0.014
2.27
-0.018 -2.23 **
High PML portfolio (dangerous: 8.36)
Constant
# of earthquakes
Estimate z -value
0.005
0.67
-0.010 -1.21
12
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Total return vs. # of quakes
Panel C:
Low financial (28%)
Constant
# of earthquakes
Estimate z -value
-0.002 -0.35
-0.003 -0.32
Low foreign (11%)
Constant
# of earthquakes
Estimate z -value
1.91
0.013
-0.014 -1.41
High financial (62%)
Constant
# of earthquakes
Estimate z -value
2.82
0.014
-0.021 -2.18 **
High foreign (37%)
Constant
# of earthquakes
Estimate z -value
0.27
0.002
-0.009 -1.25
13
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Daily abnormal return
<Large earthquakes (35)>
<Huge earthquakes (5)>
Reaction
• Gradual price drop from
day 2
• Moderate size
• Immediate price drop on day 0 and
day 1
• Large size
Recovery
• Very slow
• Not significant
• Quick from Day 2
• Significant
14
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Range of cumulative abnormal return
<5+ and 6->
PML x Fin
PML x Foreign
Coef. t
rank_pml
2
0.007
1.05
3
-0.007 -1.01
Coef. t
rank_pml
2
-0.005 -1.00
3
-0.011 -2.06
rank_fin
2
-0.007
3
-0.007
rank_foreign
2
-0.002
3
-0.001
-1.09
-1.04
-0.24
-0.21
rank_pml#rank_fin
22
-0.004 -0.50
23
-0.011 -1.28
32
0.008
0.86
33
-0.001 -0.07
rank_pml#rank_foreign
22
0.009
1.08
23
0.026
3.19
32
0.008
0.89
33
0.018
2.26
_cons
_cons
0.069
12.76
0.064
**
17.34
***
**
15
Research
Framework
Introduction
Data &
Methods
Results
Conclusions
Range of cumulative abnormal return
<6+ and 7>
PML x Fin
PML x Foreign
Coef. t
rank_pml
2
-0.009 -0.42
3
-0.020 -0.96
Coef. t
rank_pml
2
-0.026 -1.55
3
-0.015 -0.84
rank_fin
2
-0.027
3
-0.030
rank_foreign
2
-0.016
3
-0.015
-1.28
-1.49
-0.88
-0.92
rank_pml#rank_fin
22
0.000
0.01
23
0.003
0.10
32
0.035
1.28
33
0.029
1.07
rank_pml#rank_foreign
22
0.037
1.48
23
0.049
2.00
32
0.034
1.29
33
0.030
1.25
_cons
_cons
0.107
6.15
0.092
8.43
**
16
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Summary of results

Sensitivity of REIT return to gradual change in
earthquake risk perception is high
 Low PML
 High financial institution ownership

Rational price reaction after “large” earthquakes

Panic reaction after “huge” earthquakes

Sensitivity of REIT return to quick change in
earthquake risk perception is high
 High PML
 High foreign investor ownership
17
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Implications
 Probability
of rare events happening affects
asset pricing dynamics

Investors shift attention allocation to rare
event risk

Different types of investors show different
processes of attention allocation to rare event
risk
18
Introduction
Research
Framework
Data &
Methods
Results
Conclusions
Directions for future research

Longer-term effects of huge earthquakes
 Demand of tenants
 Acquisition/disposition of properties by REITs

Impacts of financial institution ownership
 Why do they respond more slowly than foreign
investors?
19