Transcript Bargaining in Real Estate Markets

Fifth Bi-National Regional Science Workshop
Tel Aviv, April 29-30, 2007
Re-Redeveloping Downtown:
Behavioral Bargaining in
Real Estate Markets
Danny Ben-Shahar
Yuming Fu
Sau Kim Lum
Tsur Somerville
Technion – Israel Institute of Technology
National University of Singapore
National University of Singapore
University of British Columbia
MOTIVATION
Explore (both theoretically and
empirically) the sale process of collectively
owned residential real estate for the
purpose of redevelopment
OBJECTIVE
• Test for determinants of successful collective
sales
• Analyze the effect of policies instated to promote
sale of buildings for redevelopment
• Look for empirical support for theories in
behavioral economics
SUMMARY OF THE RESULTS
• The greater the number of collective owners, the lower
is the probability of a sale and the longer is the sale
process
• The greater the difference between owners, the lower is
the probability of a sale and the longer is the sale
process
• The greater the redevelopment potential, the higher is
the probability of a sale and the shorter is the sale
process
THE GAME
Bargaining among
owners on the expected
price and its division
Agent believes the
expected price is
reasonable
Agent believes the
expected price is too high
Bargaining between the
owners and the buyer
We do not
observe
Bid price >
expected price
Bid price<
expected price
Bargaining among owners
on the division of the
positive surplus
Agreement
among owners
Sale
Bid price =
expected price
Bargaining among owners
on the division of the
negative surplus
Disagreement
Agreement
among owners among owners
No
Sale
No
Sale
Sale
Sale
Disagreement
among owners
No
Sale
RELEVANT LITERATURE - I
Land Assembly w/ landowner heterogeneity
Eckart (1985) and Strange (1995) assume players are
rational and predict:
– Higher land price (per sqf) for small parcels (due to
lower opportunity cost of failure)
– Land prices for assembly rise w/ number of lots to
assemble
RELEVANT LITERATURE - II
Behavioral Economics (Self-Serving Assessment of
Fairness) – Babcock and Loewenstein (1997)
– People evaluate information in ways that are beneficial to one’s own
outcome
– More likely in situations where there is ambiguity
Experimental support from the “ultimatum” game:
• Hoffman et al. (1994), Knez and Camerer (1995), and Kagel et al.
(1996)
Empirical support from union contracts:
• Babcock, Wang, and Loewenstein (1996) - strikes more likely the
larger the spread of comparable wages
• Thompson and Loewenstein (1992) - difference between
management and labour’s assessment of “fair” settlements
correlated with strike length
• Babcock and Olson (1992) - each side to a labour dispute picks a
“fair” reference focal point that is favorable to itself
THE MODEL
•
•
•
•
•
N landowners
Share of each owner is αi (
)
Owners total reservation price is B
Buyer’s offer is P
The issue is the division of P–B:
– No problem when P=B
– Problem arises with positive surpluses P>B
and negative surpluses P<B
SELF-SERVING BIAS OF “FAIR”
DIVISION
Suppose α1=33% and α2=67% then:
If P<B  β1=33% and β2=50%
If P>B  β1=50% and β2=67%
where βi is the owner’s subjective share in the
positive and negative surpluses, (P-B)
That is,
We have under-demand of negative surplus and overdemand of positive surplus
SUBJECTIVE DIVISION RULE
PREDICTIONS
• Result 1: The sum of the values requested
by the landowners increases (and thus the
likelihood of a transaction diminishes)
with the number of relatively “small”
landowners, ceteris paribus.
• Result 2: The sum of the values requested
by the landowners increases (and thus the
likelihood of a transaction diminishes)
with the variation in the ownership share,
ceteris paribus.
DESCRIPTIVE STATISTICS
A total of 285 residential projects:
Definition of Observed Variables
Mean
Std. Dev.
Binary indicator of a successful en bloc sale
0.73
0.45
Time from DGP release to en bloc sale in months. It is censored
at January 2005.
69.18
43.05
Market Information. Binary indicator for Central Area
0.80
0.40
Number of parcels in an en bloc attempt
25.88
33.80
Ownership Diffusion. The ratio of the size of the top two parcels
over that of the bottom two parcels in an en bloc attempt
1.56
0.86
Redevelopment Potential, equal to the maximum of three ratios:
(i) new maximum plot ratio to the existing plot ratio, (ii) new
height limit to the existing height limit, (iii) new maximum
developable gross floor area to existing gross floor area. It
equals 1 for cases with missing data.
3.10
2.95
Number of en bloc sale attempts (per project) made
1.25
0.53
REGRESSION RESULTS
Model
Probit
Tobit 1
Tobit 2
Tobit 3
Successful
Bloc Sale
Time to
Sale
Time to
Sale
Time to
Sale
1.356
(4.6)***
47.932
(4.3)***
41.949
(3.7)***
25.251
(1.9)*
Sale is in Central Area
0.217
(1.1)
-16.848
(2.1)**
-19.738
(2.4)**
-16.736
(2.1)**
Number of Parcels 0.5
-0.065
(2.0)**
4.772
(4.0)***
4.902
(4.1)***
3.765
(3.0)***
Ownership Diffusion
-0.543
(5.1)***
21.515
(5.1)***
25.227
(5.3)***
26.120
(5.6)***
Redevelopment Potential
0.354
(3.0)***
-13.465
(3.4)***
-14.041
(3.5)***
-15.217
(3.9)***
Dependent variable
Constant
Attempts
14.588
(2.5) **
The End