Z E M C H 2 0 1 2 I n t e r n a t i o n a l C o n f e r e n c e
126
In general, the condominium unit price is affected by differences in the unit’s
performance, including its “proprietary area” and “number of rooms,” as well as by
accessibility, including “proximity to the nearest railway station,” and the apartment
building’s performance, including its structure (
,
). The environmental performance of
the building is included in this group of attributes. The price is also influenced by the
characteristics of the surrounding environment, such as the local atmosphere and
commercial zones
(
A
k
). This is known as the “neighborhood effect.” If an analysis is
performed across a wide area covering more than one administrative ward, attention
must also be paid to differences across a wider area, such as “proximity to the central
business district (CBD)” and differences in administrative services available at the
regional level. Many prior studies have indicated that there are also disparities by railway
line, along which some cities in Japan have been developed. Therefore, attention is also
paid to differences in environmental features across a wider area than the
neighbourhood (
C
l
)
.
3.2. Demonstration analysis-Estimation model-
For hedonic function calculation purposes, we estimate two demonstration models,
Models 1 and 2. As explained above, the data includes asking and transaction samples.
This study stacks them and estimates the difference between asking and transaction
samples using a transaction dummy (
TrD
i
). In other words, we apply the hedonic price
formula to the 82,270 samples (the sum of asking and transaction samples), while at the
same time estimating the disparity between transaction and asking samples, using the
2,063 samples. It should be noted, however, that because of the limited number of
transaction samples, this estimation is not accurate enough and statistically significant
results are difficult to obtain.
The effect of green labels ( ), which is the subject of this research, is estimated
using several different types of dummy variables for green labels. First, we estimate the
average effect of green labels independently of the environmental performance level and
the time of transaction, then estimate how the effect of green labels changes with the
time of transaction, using the cross term of the environmental performance level and the
time of transaction. Next, we estimate how the effect of green labels changes with the
type and positiveness of green features, using the ratings in each environmental
performance assessment item. Finally, even if sellers ask for higher prices because of
green labels or environmental performance offered, buyers or consumers may not
accept the price differences and the final transaction prices may be almost the same as
those for condominiums without green labels. By adding a cross term of green label ( )
and transaction dummy (
TrD
i
), we can estimate both of them.
First, in Model 1, we analyze price gaps between condominiums with and without
green labels under the Tokyo program, using a dummy for the environmental variable
( ) that shows whether or not these labels are obtained. The possibility of the effect of
green labels being different between asking and transaction samples is represented by a
cross term. The function form is a semi-logarithm, and square terms are added for some
variables, such as the size of the building. This is intended to deliberately exclude the
possibility of an alternative variable for building size being created because green labels
are required for buildings whose total floor area exceeds 10,000 square meters. For
explicated variables, the natural logarithm of the price per square meter of condominium
unit is used. The results, therefore, provide a rough estimate of the percentage by which
the price changes along with each variable.
