3
1.1 Research background Every seller wants as higher selling price and as shorter time on
market as possible. However, according to the theoretical framework
of Simon (1995) and later theories of Wheaton (1990), Yavas (1992),
Krainer & LeRoy (2002), Anglin & et al (2003), Lin & Vandell
(2007), Cheng & et al (2008), these are two conflicting objectives.
That means home sellers always have to trade between these two
objectives. Accordingly, as home sellers spend more time searching
for potential buyers, they will have more likelihood of finding more
quality buyers who are willing to pay a higher price for the forsale
property. In this relationship, the selling price acts as a tool to
determine the tradeoff of the seller, a high listing price is a seller's
signal of a high selling threshold and thus limiting the number of
buyers. This leads to a longer listing period, but the selling price will
be higher due to better quality of potential buyers (Hoeberichts et al.,
2008).
However, in particular, 8/18 studies compiled by Sirmans et al.
(2005) found a relationship contrary to the above theory between
selling price and forsale time. Similarly, Johnson et al. (2008)
synthesized 108 relevant studies during the period of 19952007,
only 29 studies had the right results in theoretical relationships, 52
studies found the inverse relationship, and 24 cases found no
relationship between these two factors. Benefield et al. (2014)
summarized 197 related studies and found similar findings with 100
studies identifying the theoretical inverse relationship.
In order to explain the inverse correlation between selling price and
time on market explored by emprirical studies, basing on retail
model of Lazear (1986), Taylor (1999) has developed a theoretical
4
framework of market discrimination. Accordingly, when the quality
of housing is difficult to be observed by buyers, a house with a long
forsale period will be a signal of having "bad" quality and therefore
the transaction price of the house will decrease.
The relationship between the price and time on market is both
theoretically and empirically ambigious, so the role of the listing
prices in transaction prices and time on market has not yet been
agreed among researchers. Therefore, many researchers suggest that
more empirical studies should be conducted on the impact of factors
on the relationship between the selling price and time on marrket,
and the probability of selling in different real estate markets,
especially newly developed and smallsized housing markets, to
contribute to raising awareness and improving the efficiency of these
markets (McGreal, Adair & Brown, 2009; Filippova & Fu, 2011;
Cirman et al., 2015).
In particular, there is only a small number of research topics on the
sale ability of house around the world, such as measure the effect of
selling strategy (Kluger & Miller, 1990; Hui et al., 2012); atypical
degree of the house (Krainer, 1999); the value of the house (Smith,
2010); sellers’ motivation (Johnson et al., 2008; Cirman et al., 2015)
on probability of sell of a house. Although having found to have an
impact on the probability of sale of the house, it is limited to
measuring the static impacts of these factors. However, the longer
the time on market, the more likely homebuyers’ behavious will
change (Taylor, 1999) so according to the author, the impact of
factors on the probability of sale will varies accordingly. However,
this issue has not yet mentioned on prior studies. This is the research
gap need to be considered.
5
In addition, the housing studies in Vietnam mostly focus on the
impact of factors on house prices (Kim, 2004, 2007; Bui, 2020a,
2020b; Seo and Kwon, 2017). The research in buyerseller behavior
strategy, the liquidity and the selling ability of the house are not
available in Vietnam.
Therefore, within this thesis, regarding home sellers, the author
conducts research on the sellers' behavior, in particular, the role of
listing price strategy in selling price, time on market, and the
probability of sale in the detached housing market in Ho Chi Minh
City. In particular, the impact of the listing price strategy on the
probability of sale of a house will be measured against the different
time on market (1 month, 3 months, 6 months, and 9 months) to
determine the fluctuations of this effect over time on market.
In homebuyers site, a finding of the author in understanding the
theoretical frameworks about current homebuyers' behavior is that
there is no theoretical framework that analyzes the impact of their
current houses’ characteristics (houses where they are residing in) on
their current housing search behavior. In particular, a buyer with
experience in some charecteristics of the prior house such as the
the heat, the stuffiness of the house suffered by the sunshine in the
afternoon will concern more about the view of the new house.
However, there are no researches mentioned about these issues both
in empirical and theory. Therefore, in this thesis, the author also
conducts a theoretical framework to analyze the impact of
homebuyers’ prior houses on their current search behavior. 1.2 Research Objectives With the aforementioned subjects, the author conducts research on
two aspects of the housing market in HCM city: analysis of housing
6
supply and analysis of housing demand. On the supply side, the
author analyzes the seller's bidding strategy. On the demand side,
this thesis focuses on house buying behavior under the influence of
current house’s characteristics. Specifically:
Analyze the relationship between the listing price strategy of the
seller and the selling price, the time on market, and the probability of
sale corresponding to different periods.
Use the search behavior theory to develop a theoretical model that
analyzes the impact of characteristics of the buyer's current home on
their buying behavior and tests this correlation by empirical results. 1.3 Research methods Two different research methods will be applied to the two research
objectives.
With the aim of analyzing the relationship between the listing price
strategy and the selling price, the time on market, and the probability
of sale corresponding to different periods. The author will apply the
3step research method. In particular, step 1 is be to determine the
listing price strategy based on the difference between the actual
selling price and the market price of the house, which is estimated
from the properties of the house using hedonic model. Step 2
measures the effect of this listing price strategy on the transaction
price and time on market through quantitative models. Step 3
measures the impact of the placement strategy on the probability of
sale through the Cox Proportional Hazard Model with 1, 3, 6, and 9
month sales timeline, which is to consider changes in the influence
of these factors over these sales time points.
With the objective of developing a theoretical framework analyzing
the influence of old houses’ properties on the current housing search
7
behavior of homebuyers, the author will outline the theory related to
search behavious of current home buyers, then upholding and
developing their theoretical framework. The conclusions of the
theoretical framework will be tested experimentally to determine the
validity of the developed theoretical framework.
1.4 Thesis Structure Follow this chapter are chapter 2: Theoretical basis; Chapter 3:
Results of measuring the impact of pricing strategies; Chapter 4:
Develop a theoretical framework analyzing the influence of
homebuyers’ current dwellings on their behavior; Chapter 5:
Conclusions and recommendations
Chapter 2: Theoretical basis
2.1 Some related concepts In this content, the author presents a number of related concepts:
such as detached houses, prices related to housing, listing price
strategies including underpricing and overpricing strategies.
2.2 Theoreical basis In this content, the author presents the following 3 groups of
theoretical bases:
2.2.1 The group of theories related to the relationship between
listing prices, selling prices and time on market on the housing
market This consists of theoretical basis on the relationship between price
and time on market by Cheng et al. (2008), framework theory of
psychological stigma of Taylor (1999), theoretical model of fishing
behavior by Sun and Seiler (2013). Including:
8
* The theoretical basis for the relationship between housing price
and time on market by Cheng et al. (2008) was developed based on
the theories of Wheaton (1990), Krainer & LeRoy (2002), Anglin &
ự ồ đ ng s (2003), Lin&Vandell (2007). On the basis of the theory by
Cheng et al. (2008), a home seller would wait for n home buyers to
come and offer a price, and then the buyer would negotiate with the
buyer who offers the highest price to sell the house. And then it is
the problem of the seller to dread the risk of long time waiting, that
is, to determine the optimal n*, to achieve the highest level of risk
adjusted selling price expectations, in an ideal case when the buyer
does not leave the market, and even if there is a percentage of the
buyer leaving the market. As a result of the theoretical framework,
Cheng et al. (2008) identified a positive relationship between long
sales and high expected riskadjusted trading prices in both cases.
* According to the theoretical framework of Taylor's discriminatory
psychology (1999), the process of selling a house consists of 2
phases, buyers in each phase will bid together, the winning bidder
will review the house and decide based on their judgement, knowing
that the judgement always entails a rate of error. Theoretical results
show that, when buyers in phase 2 have not been reviewed in phase
1, probability of ending up with good houses in phrase 2 is always
lower than in phrase 1. Therefore, according to Taylor (1999), longer
listing time is a signal of the bad quality of the house (called stigma
9
signal), so the relationship between transaction price and time on
market is negative.
* The theory of the fishing behavior of sellers by Sun & Seiler
(2013) is also a process of selling a house in 2 phases with limited
time for sale, so in the final stage, the seller has to sell at the average
price market, and therefore in phase 1 the suggested price x must be
greater than the market average to be considered. Since then Sun &
Seiler (2013) determines the seller's maximum acceptable x* price
and determines the fishing behavior is the setting of the above
acceptable minimum price compared to the market average price of
the house. The theoretical results of Sun & Seiler show that in most
cases, sellers always have a motive for fishing, and the higher the
quality of the house (the greater the value to the seller), the greater
the fishing behavior will be showed by sellers.
From these three theories, the author of the thesis proposes two
research hypotheses that need to be clarified:
H0: An overlisting price strategy will signal a high selling threshold and thus increase the selling price of a house but at the same time
prolong the time on market, derived from the theory of Cheng et al.
(2008).
H1: An underlisting price stategy will be a signal to buyers about the possibility of a problem in the house (derived from the theory of Sun
& Seiler, 2013), and therefore the house will become difficult to sell
at lower prices and with longer time on market (derived from
Taylor's theory, 1999).
2.2.3 Theory of the Cox home sale ability model
10
To measure the impact on the likelihood of a risk (probability of
sale), in the past, researchers used the Survival model to measure the
extent of the impact of false factors that were deviated from the
baseline rule of the study subjects, but the problem is that this basic
rule of life is not observable, which limits the applicability of the
survival time model. However, Cox (1972) has developed into a Cox
risk model with the advantage of not needing to know the basic rules
of risk due to the comparison with the standard subjects of the
sample instead of the rules. Basically, this has expanded the
applicability of the Cox risk model (Cajias & Freudenreich, 2018).
Therefore, the dissertation will also apply the Cox model to measure
the impact on probability of sale of a house (should be called the
Cox sale model), and establish the Cox model with timelines of 1, 3,
6, 9 months to analyze the fluctuations of these effects over time.
Cox model will estimate the HR ratio (hazard ratio) of the
explanatory variables in the model and thus tell us whether that
variable has an effect of increasing (HR> 1) or decreasing (HR <1)
the probability of sale of the house, and the impact rate is determined
to be (HR 1)% when increasing and (1 HR)% when reducing the
likelihood of selling.
2.2.4 Group of theories related to search behavior on the housing
market With the aim of developing the theory of the influence of
homebuyers’ current dwellings on their behavior, in this section, the
author also presents the theoretical basis related to the search
behavior of homebuyers as the theory of efficient buyers' behavior
(Cronin, 1982), the theory of the influence of information over time
(Turnbull & Sirmans, 1993) the influence of social capital (Tu et al.,
11
2016), and the effect of spatial information (Qiu & Zhao, 2018) on
buyer search behavior. Survey results show that, basically, the
reasonable search behavior of home buyers is the decision between
continuing to find or buying a home under consideration based on
comparing the benefits between these two options, and the search
stops are determined when the benefits of continued search are equal
to the benefits of buying a home, and the benefit level at this stop is
called the threshold interest (Cronin, 1982). The recent theories of
Turnbull & Sirmans (1993), Tu et al. (2013), Qiu & Zhao (2018) are
all developed based on the transformation/expansion of an element
in the benefits of continued search, and analyze how this change will
affect homebuyers' threshold benefits.
On this basis, when developing the theoretical framework of the
influence of the old house on buyer behavior, the author will add the
advantage that the buyer benefits from the old house to the continued
search and then analyze the effect of the benefits of their current
houses on the threshold benefit at the stopseraching point.
2.3 Overview of related studies In this content, the author reviews studies related to the relationship
between the listing price, selling price and time on market in the first
stage (Cubbins 1974, Belkin et al. 1976, Miller 1978) and expands to
the recent research branches of this issue. In addition, the author also
examines studies related to the probability of sale of a house.
Chapter 3: The impact of listing price strategies
3.1 Analytical models and research data
3.1.1 Econometric model
To measure the impact of the listing price strategy on the selling
price, time on market, and probability of sale, the author will apply
12
the method suggested by Kluger & Miller (1990) and was applied in
the research of Krainer (1999), Johnson & et al (2008), Smith (2010),
Hui & et al (2012), and Cirman & et al (2015). In this research
methodology, there are 3 steps as follows:
Stage 1: Hedonic model was applied to estimate the expected market
price of the house basing on housing charecteristics of the house,
then determining pricing strategy. In this stage, the author estimate
expected market hedonic model as follows:
is the natural logarithmic value of the real selling price of house j in
the market (in million dong), therefore, is the expected market price
of house j estimated from the properties Sj (structural property), Lj
(position and accessibility), Nj (property of the surrounding area) ),
ε and stands for error.
To get the best result, some diferent hedonic models estimated as
equation 1, the author chooses the best model to estimate the expected
market price of the house, and from that calculate the DOP price
difference, representing the seller's pricing strategy as follows:
Stage 2: Measure the impact of pricing strategy on selling prices and
time on market. The pricing strategy (DOP) of the houses will be
used to measure the impact on selling price by below equation:
However, due to the potential proximity between DOPj and , the
author will replace DOPj with the dummy variable Dum_DOP j with
the value 1 representing the overlisting price strategy (DOP> 1) and
0 representing the underlisting price strategy (DOP <1) and
equation (2) is now rewrited (2').
13
+ And measure the impact on time on market through the equation:
(3)
Sateg 3: Measure the impact of pricing strategy on probability of sale
of a house through the equation:
In particular: is the probability of sale of the house j with the
property conditions Sj, Lj, Nj and the DOPj and listing price
strategies are the probability of sale of the standard house with the
property conditions of the standard house S0, L0, N0 and DOP0. Then, the dependent variables PS and TOM and the explanatory variables
Sj, Lj, Nj of the house, used in equations (1), (2), (3), (4), are
discussed in section 3.1.1.1 and 3.1.1.2.
3.1.2 Research data
The author surveyed 460 individual housing transactions in the urban
area of Ho Chi Minh City based on the probability of stratification
by geographic area, the districts with active transactions will be
allocated a large number of samples and vice versa. Subjects of the
survey are separate housing transactions (excluding villas) in the
secondary market with the participation of brokers to ensure (1) the
negotiated position between buyers and sellers is equivalent
(secondary market), (2) sellers obtain information about average
prices of similar houses in the area (broker) when making decisions
and listing prices, (3) the similarity in type of house and land use
purpose (individual houses in urban areas and excluding villas) to
reduce the variance change.
14
In particular, data on prices, time on market, structural and location
features were asked by brokers, while accessibility and surroundings
were asked by homebuyers (also information about their old houses).
Table 3.2: Descriptive statistics of data Description
Price Time on market Age of house Lot size Square footage
Varia. name Price Tom Age Slotarea Floorarea Outside
Mean Std. Dev. Min Max 26 600 884 30 320 600 5
7 932.45 5 777.35 130.14 114.38 7.18 8.93 36.64 71.21 111.70 186.75 1.293 2.531
900 1 1 25 44 1
Shape
shape of the land
0.82
0.39
0
1
Wide Long Nbedr Nbathr
4.49 15.66 3.83 3.50
1.44 4.93 1.88 2.00
3 6.8 1 1
12 32 14 15
Sun
0.36
0.48
0
1
Face Dstreet
Unit Million VND Days Years square meter square meter 1 – oldest; 5 – newest Outside charecteristics 0 square/grew bigger 1 otherwise meter meter No. of rooms No. of rooms 0 sunshine 1 otherwise 0 alley/ 1 frontage meter
0.23 89.18
0.42 163.39
0 0
1 1 000
Widestreet meter
8.69
6.30
1
30
Dcbd Tcbd Tworkpla
8.15 22.98 14.38
3.92 10.22 8.00
0.6 1 1
16.4 60 40
Safe
5.83
1.21
1
7
Waste
0.61
0.37
0
1
Smelly
4.68
1.95
1
7
Noisy
5.06
1.81
1
7
Flooding
0.90
0.29
0
1
Width of house Long of house Number of bedrooms Number of bathrooms house facing afternoon sunshine (hot) Road frontage Distance from frontage width of the road in front of the house Distance to CBD Time to CBD Time to working place Status of safety in neighbourhood Status of garbage collection system Status of smell in neighbourhood Status of quiet condition in neighbourhood Status of flooding in neighbourhood
Kilometer Minutes Minutes 1 is worst 7 is best 0 no collection system. 1 otherwise 1 is worst 7 is best 1 is worst 7 is best 0 no flooding 1 flooding
(Source: Research calculations from selfsurvey data)
3.2 Identifying pricing strategies
3.2.1 Research methodology to identify pricing strategies This is the first step in the 3step method of the thesis to measure the
impact of the listing price strategy with the objective of indentifying
15
the DOP listing price strategy for the house. The use of a direct
difference between the listing price and the actual trading price will
cause potential deviations due to the concurrency (stick) between
these two variables (Yavas & Yang, 1995; Hui & co., 2012), so the
author applied the method proposed by Kluger & Miller (1990) and
successfully applied by related studies, which is an estimate of the
expected market price of the house through the hedonic model and
use this price range to calculate the DOP. Therefore, in this step, the
author will establish a hedonic model in the form of a model (1) and
use it to estimate the market's expected market price, and then use
this price to calculate the DOP deviation compared to the actual
listing price.
3.2.2 The results of the seller's listing price strategy model
Model 1
Model 2 Model 3
Variable s
Coef.
VIF
Coef.
VIF
Coef.
VIF
Coef.
VIF
Model 4 Robust Std. Err.
0.0318
Robust Std. Err. 0.0202
3.4
Robust Std. Err. 0.0417** 0.0201
3.6
Robust Std. Err. 0.0419** 0.0179
3.5
3.7
0.0374** 0.0178 0.1522** *
0.0374
0.1818***
0.0514
4
0.163***
0.0391
2.3
0.1551*** 0.0373
2.3
2.3
1.26E07
3.7E07
3.7
0.6044***
0.0741
5.1
5.1
0.0715
0.5297** *
0.0556
2.8
0.0559
2.8
0.5261** *
Lnage lnfloorar ea floorare asqu lnslotare a slotareas qu Face
3.5E06* 2.1E06 0.0631 0.1056*
2.9 2.8
2.9 3.3
2.2E06 0.0584
0.0892**
0.0348
1.2
1.2
0.0334
0.0894*
0.0346
1.2
0.0855** 0.0348
1.2
0.0153***
0.0041
2.6
2.8 1.3 1.8
0.6253** * 3.9E 06* 0.0805 0.0843** 0.0154** 0.0037 * 0.1821** 0.0919 1.5E04* 8.5E05
0.0021 0.018*** 0.1466* 0.0834 1.5E04* 7.9E05
1.5 1.3 1.5
0.0181*** 0.1456* 1.5E04*
0.002 0.0818 8E05
1.5 1.3 1.6
Shape widestre et Acar dstreet lntwork pla
0.021
1.2
0.0471** 0.0213
1.3
0.0488** 0.0212
1.3
Lntcbd
1.8
0.0247
0.067**
0.0296
1.9
0.0638** 0.0298
1.9
1.2
0.0509** 0.0694** * 0.0672** *
0.0242
Sun Safe Waste
0.024 0.0137 0.0447
1.3 1.7 1.2
0.0241 0.0138 0.0449
1.3 1.7 1.2
Smelly Noisy
0.0543** 0.0143 0.0585 0.0362** *
0.0094 0.0105
2 1.9
0.0565** 0.0146 0.0568 0.0363** *
0.0094 0.0105
2 2
Table 3.3: Estimation results of house price models
16
0.0294** *
0.0301** * 0.0904**
0.0425
1.3
0.0878** 0.0428
1.7
0.063
1.7
flooding slig_flo od stri_floo d
0.1957*** 5.4636** *
0.2621
4.8235***
0.3116
5.0776** *
0.2929
5.4684** *
0.2605
Yes
Yes
Yes
Yes
_cons D. C.Dum my
0.8706 0
0.8795 0
0.888 0
0.8887 0
0.24958
0.24195
0.23412
0.23364
lnprice
lnprice
lnprice
Lnprice
R squared Prob(F) Root MSE Dep. Var. N. of obs
448
448
448
448
Note: The models in the table are estimated by the method of OLS with strong standard errors.. *, **, *** respectively represent significance levels at 10%, 5% and 1% respectively Checking VIF variables in the model without multicollinearity signs.
(Source: Estimates based on survey data of the study)
Excluding a number of variables with multicollinearity, model 4
results will be used to estimate the expected market price based on
the properties of the house because of the higher degree of
interpretation and accuracy. The difference between the actual listing
price and the expected market price will then represent the sellers'
pricing strategy.
3.3. Measuring the effect of the listing price strategy on the price
and time on market
3.3.1 The research method applied in this step This is the second step in the 3step method, the goal of this step is to
measure the impact of the listing price strategy on house prices and
time on market. First, the author applies model 4 in table 3.3 to
estimate the expected market price of housing (variable Pricef). And
the DOPj house listing price strategy will be determined as follows:
17
The DOPj house listing price strategy will then be used as an
explanatory variable in equations (2 ') and (3) to measure the impact
of the listing price strategy on transaction prices and time on market
of the house to answer two hypotheses H0 and H1 of the thesis.
Particularly for equation (2 ') due to limiting the potential proximity s, the author will replace DOPj with the dummy between DOPj and Pi
variable Dum_DOPj.
3.3.2 Results of measuring the impact of a listing price strategy
on the selling price As mention above, the listing price strategy (DOP) in this case will
be change to dummy variable Dum_DOP.
Table 3.4: Results of measuring the impact of listing price strategies on housing prices
Variable s
Coef.
VIF
Coef.
VIF
Coef.
VIF
Coef.
VIF
Model 5 Model 6 Model 7 Robust Std. Err.
Robust Std. Err.
Robust Std. Err.
Model 8 Robust Std. Err.
0.0218
1.06
0.3187*** 0.0206
1.08
0.3513*** 0.0189
1.1
0.0186
1.1
Dum_D OP
0.0176
3.4
0.0651*** 0.0167
3.5
0.0618*** 0.0143
3.6
0.0143
0.3553** * 0.0692** *
3.7
0.0383
4.0
0.1287*** 0.0276
2.3
0.1037*** 0.0258
2.3
0.1078*** 0.0255
0.3053** * 0.0505** * 0.1696** *
2.3
3.7
0.0506
5.1
0.5514*** 0.0348
2.8
0.0349
2.8
5.35E08 2.3E07 0.6055** *
0.0556
5.1
0.5458** *
2.9 2.8
1.2E06 0.0455
2.9 3.4
lnage lnfloorar ea floorarea squ lnslotare a slotareas qu face
0.6226** * 2.5E 06** 0.0598
2.11E06 1.3E06 0.1034* 0.0533
0.1082*** 0.0267
1.2
0.1008*** 0.0241
1.2
0.1012*** 0.0218
1.2
0.0951*** 0.0221
1.2
0.0167*** 0.0032
2.6
0.0183*** 0.0026 0.0667
2.8 1.3
0.0013 0.0595
0.0201*** 0.0012 0.054
1.5 1.3
1.5 1.3
shape widestre et acar
5.9E05
1.8
4.7E05
0.0911* 1.6E 04***
4.7E05
1.5
1.6
dstreet
0.02*** 0.0934 1.7E 04*** 0.0645** *
0.0165
1.3
0.0153
1.3
0.015
1.3
lntworkp la
0.1301* 1.9E 04*** 0.0692** * 0.0987** *
0.0675*** 0.0954** *
0.0169 0.0733*** 0.0178
1.8 1.2
1.9 1.3
1.9 1.3
lntcbd sun
0.0094 0.03
0.0095 0.0301
1.7 1.2
1.7 1.2
safe waste
0.0067 0.0072
0.0067 0.0071
2 1.9
2 1.9
smelly noisy
0.1003*** 0.0179 0.0577*** 0.0159 0.0299** * 0.0788*** 0.0467** * 0.0385**
0.0178 0.0612*** 0.0156 0.0307** * 0.0763** 0.0469** * 0.0375***
18
*
0.1449*** 0.0234
1.3
0.1414*** 0.0236
1.7
flooding slig_floo d stri_floo d
0.3124*** 0.0391
1.7
4.7527** *
5.6097** *
5.6037*** 0.1668
0.2352
5.3157*** 0.2003
0.1648
Yes
Yes
Yes
Yes
0.917
0.9294
0.9471
0.949
0
0
0
0
0.2001
0.18546
0.16104
0.15844
cons D.C.Du mmy R squared Prob(F) Root MSE Dep. V. N. of obs
Lnprice 448
lnprice 448
lnprice 448
lnprice 448
Notes: The models in the table are estimated by the method of OLS with strong standard errors. *, **, *** respectively represent significance levels at 10%, 5% and 1% respectively. Checking VIF variables in the model without multicollinearity signs.
(Source : Estimates based on survey data of the study)
Estimated results show that the overlisting price strategy has a
significant impact at 1% with the level of helping increase 35% of
the transaction price of the house compared to the strategy of under
listing. In addition to the listing price trategy, a number of other
properties of a home have been identified as having an influence on
the selling price of a property such as age, area, type of land,
proximity to the center, workplace, accessibility and safe and
hygienic surroundings.
3.3.3 Results of measuring the impact of listing price strategies
Table 3.5: Results of measuring the impact of the listing price strategy on time on market
Model 11
Variables
Model 9 Model 10 Coef. Robust Std. Err. VIF 1.05 0.2856 3.39 0.1149 3.97 0.2295 3.71 1.4E06 2.76 0.2234 2.78 0.2257 1.19 0.1746 2.56 0.015
0.4711* DOP Lnage 0.1009 Lnfloorarea 0.5771** floorareasqu 2.57E06* 0.6558*** Lnslotarea Slotareasqu 0.2386 Face 0.075 Shape 0.0114 Widestreet Acar Dstreet Lntworkpla Lntcbd Sun
Coef. Robust Std. Err. VIF 1.05 0.2817 0.1171 3.5 2.24 0.1572 2.7 0.2231 3.34 0.244 1.22 0.1672 2.81 0.0158 1.28 0.256 1.77 0.00039 1.24 0.1048 1.84 0.1369 1.24 0.1423
0.4849* 0.0731 0.2536 0.6703*** 0.5079** 0.0322 0.025 0.0811 9.91E04** 0.0693 0.4075*** 0.0727
Coef. Robust Std. Err. VIF 1.04 0.2646 0.449* 3.58 0.1096 0.1184 2.31 0.1485 0.1785 2.83 0.2312 0.8452*** 1.24 0.1693 0.0319 1.49 0.0117 0.0028 1.28 0.2882 0.2894 1.54 0.00038 5.75E04 1.27 0.1063 0.0123 1.88 0.1411 0.3694*** 1.27 0.1446 0.1137
on time on market
19
1.0985
1.4167
0.1885*** 0.1717 0.1144*** 0.1019** 0.2914 1.9464
0.0576 0.1537 0.0416 0.0462 0.233 1.4572
1.65 1.17 1.98 1.94 1.27
4.2946***
1.4643
Yes
Yes
Yes
0.3133 0 1.194 lntom 448
0.2601 0 1.2277 Lntom 448
0.2803 0 1.2165 Lntom 448
Safe Waste Smelly Noisy Flooding slig_flood stri_flood _cons D.C.Dumm y Rsquared Prob(F) Root MSE Dep. Var. N. of obs Notes: The models in the table are estimated by the method of OLS with strong standard errors.. *, **, *** respectively represent significance levels at 10%, 5% and 1% respectively. Checking VIF variables in the model without multicolline arity signs.
(Source: Estimates based on survey data of the study)
The significance level of time on market models that fluctuates
between 26% 31% is consistent with other authors' timeforsale
models. The house's listing price strategy (DOP) was found to have a
10% meaningful impact which shortened the time on market of the
home. Thus, along with Model 8, the author concludes that we need
to reject the hypothesis H0 and accept H1. Accordingly, a high
bidding strategy acts as a signal of the "good" quality of the house,
and therefore the buyer not only accepts to pay a higher price, but
also has an incentive to buy faster.
20
The reason is due to the shortage of housing in the market in the
research period, with the absorption rate increasing up to 50% 59%.
In addition to the listing price strategy, small houses, near the center,
in a secure area and good environment are also factors that attract
buyers and help shorten the time on market of the house.
Hình 3.3 The house price index in HCM
ồ Ngu n: Savills Research and Consultancy
3.4 Measure the impact of the listing price strategy on
probability of sale of a house
3.4.1 Research method applied in this step Cox saleability model is developed from the viability model, so there
are two main components: the ability to survive S(t) and the
likelihood of risks h(t). Where S(t) is the likelihood that a house still
exists in the market at time (t) and h(t) is the likelihood of a risk
occurring at time t. So, the Cox model requires some changing
observations (censored observations) and others do not change the
status (censoring observations) between before and after time t.
Therefore, to solve this problem, the author breaks the study time
into several timelines: 1, 3, 6, 9 months (time points t). And for each
21
timeline, the censorimg variable will get a value of 1 for homes that
have a shorter listing time, and vice versa. This means we have 4
moderating variables corresponding to 4 timelines: onemonth,
threemonths, sixmonths, ninemonths. The Cox models measure the
effect of the listing price strategy and other properties on the
probability of sale with the dependent variable in the models being
the time on marekt under the same censorship conditions.
3.4.2 The results measuring the impact of listing price strategy
on the probability of sale
Table 3.7: Estimation results of saleability of Cox model For forsale times of 1 month, 3 months, 6 months, 9 months Breslow method for ties
1 month Cox model
6 month Cox model
Haz. Ratio
Haz. Ratio
Haz. Ratio
Haz. Ratio
DOP Age Lnfloorarea lnslotarea widestreet Acar Dstreet Lntcbd lntworkpla Sun Safe Waste Smelly Noisy Flooding
Robust Std. Err. 0.4264 0.0111 0.1700 0.1439 0.0119 0.2476 0.0003 0.1904 0.1177 0.1586 0.0641 0.2191 0.0523 0.0328 0.2175
9monthCox model Robust Std. Err. 0.3829 0.0096 0.1755 0.1240 0.0110 0.2255 0.0004 0.1737 0.1165 0.1333 0.0603 0.2306 0.0403 0.0322 0.2192
1.4224 1.0258 1.1882 0.5856** 0.9945 0.9281 1 1.1121 1.1699 1.1336 1.1359** 1.6146*** 1.0802** 0.892*** 1.0569
1.5549* 1.0243 1.2188 0.5898** 0.9909 0.8772 1 1.2398 1.1726 1.107 1.1472*** 1.6931*** 1.0811** 0.8762*** 1.0754
Robust Std. Err. 0.3293 0.0095 0.1632 0.1236 0.0097 0.2506 0.0004 0.1366 0.1139 0.1333 0.0603 0.2317 0.0398 0.0332 0.2203
No. of subjects = 448
No. of failures = 380
Time at risk = 51242
Wald chi2(15) = 56.09
Prob > chi2 = 0
2.0827* 1.0106 1.207 0.3305*** 0.9829 3.406* 1.0007 0.9832 0.9238 1.1939 1.1533* 1.6523*** 1.2283*** 0.829*** 1.1224 No. of subjects = 448 No. of failures = 108 Time at risk = 51242 Wald chi2(15) = 60.32 Prob > chi2 = 0 Log pseudollh = 626.55
3 month Cox model Robust Std. Err. 0.7966 0.0176 0.3023 0.1228 0.0212 2.4489 0.0005 0.2434 0.1797 0.2527 0.0906 0.2781 0.0974 0.0500 0.4305 No. of subjects = 448 No. of failures = 260 Time at risk = 51242 Wald chi2(15) = 91.65 Prob > chi2 = 0 Log pseudollh = 1469.78
1.645* 1.0107 1.0723 0.622** 1.0023 0.8632 1.0005 1.1744 1.0007 1.1703 1.1541* 1.7565*** 1.153*** 0.8402*** 0.9677 No. of subjects = 448 No. of failures = 366 Time at risk = 51242 Wald chi2(15) = 62.43 Prob > chi2 = 0 Log pseudollh = 1993.93
Log pseudollh = 2056.96
Notes: The models in the table are estimated by the method of OLS with strong standard errors. *, **, *** respectively represent significance levels at 10%, 5% and 1% respectively. Checking VIF variables in the model without multicollinearity signs
22
(Source: Estimates based on survey data of the study)
The results show that the listing price strategy has a strong impact on
the probability of sale of houses in the first 30 days of sale, after that,
this effect dimimishes and after 180 days of sale, there is no effect.
This shows that the overlisting price strategy is a "good" signal of
the quality of the house to the buyer, which increases the probability
of sale of the house, but when the time for sale is prolonged, it is a
good signal. “not good” effect on house quality (Taylor, 1999) and
thus counteract the impact of the listing price strategy.
The same problem was found for the house area factor. A small
singlefamily house always attracts buyers (due to budgetary issues)
so it is likely to sell out in the first 30 days of sale, but then the time
for resale is a bad signal which wore out impact of this factor.
The ability of cars entering the house is different, according to the
author, is due to the difference between the two groups of buyers, the
buyer has a financial surplus and those with financial constraints.
Chapter 4: Developing a theoretical framework that analyzes the
impact of homebuyers’ current dwellings on their behavior
4.1 Developing theoretical framework Each house listing in the market conclude a vector of charecteristics,
Xi, then give the buyer the utility u(xi) and it has the value for buyer
ự is with the price Pi. If buying, the buyer gains the benefit , knowing Gi ∈ [ ∞, + ∞] and follow the distribution rule F(G). (Cronin, 1982; ồ Turnbull&Sirmans,1993; Tu & đ ng s , 2016; Qiu&Zhao, 2018). If
you do not buy and continue searching, the buyer can still benefit
from the old house with the vector of charecteristics x0 and gains the
benefit from the old house during the search, the costs come from
23
finding SC and expecting the benefit of the house to find E (G), i.e.
The benefit of continued search is + E (G) SC.
And the buyer makes the decision is to Max [Gi, + E (G) SC]. If
the house i is considered to have Gi < + E (G) SC, the buyer will
continue seeking for another house. If Gi + E (G) SC, the buyer
will decide to buy the house. Therefore, the minimum benefit level
of the house i for the buyer to buy, also called the G* purchase
threshold, is determined by: G* Gi = + E (G) – SC. With:
After some necessary changes, we have:
The above equation shows the optimal stop point. Comparing
optimal stop point equation in related theories, this stop rule is
affected by the charecteristics of the old house. Then the author
conlude 2 propositions as follows:
Proposition 1. The relationship between purchase threshold and the
charecteristics of the old house is determined:
The affection of the characteristic of the old house on seeking
behavior depend on the role of these charecteristics for the old house. Proposition 2. The trend in which the old home property's charecteristics tends to seeking time T of a home buyer depends on its role in the old home:
Is shows that an unfavorable feature of an old home will have the
effect of shortening the search time of a home buyer.
Thus, the analysis results of the theoretical model show that the
properties of current dwelling (expressed in its benefit value, G0)
have an impact on the buyer's current buying behavior.
24
4.2 Empirically test the impact of current dwellings
4.2.1 Experimental testing method To test the theoretical impact of old home properties on the present
searching behavior of homebuyers, the author will use a hedonic
model to measure the effect of urban flooding of the prior house
(Oldflood) on the transaction price and the time on market of the
present choiced house of buyer.
4.2.2 Experimental test results about the influence of
homebuyers’ current dwellings on their current buying behavior Table 4.1: Estimation results of the prior house’s flooding feature on the transaction price and time on market of present house Time on market model House price model Model 12 Model 13 Coef. Coef. Coef. Variables
0.1208 0.1864 0.8548*** 0.0266 0.1065 0.0012 0.0006 0.006 0.3738*** 0.183***
Robust Std. Err. 0.018 0.036 0.056 0.035 0.024 0.002 0.000 0.022 0.030 0.014 0.045 0.010 0.011 0.042 0.115*** 0.1027** 0.2911 LnAge LnFloorarea LnSlotarea Shape Sun Widestreet Dstreet LnTworkpla LnTcbd Safe Waste Smelly Noisy Flooding Oldflood 0.0387** 0.1609*** 0.5282*** 0.0864** 0.0539** 0.0189*** 0.0002** 0.043** 0.0709** 0.0129 0.0632 0.0357*** 0.0315*** 0.0957** 0.0404** 0.169*** 0.5261*** 0.0849** 0.0564** 0.019*** 0.0002** 0.0439** 0.069** 0.0125 0.0643 0.0365*** 0.0303*** 0.1021** 0.0631*** Robust Std. Err. 0.018 0.036 0.055 0.035 0.024 0.002 0.000 0.022 0.029 0.014 0.045 0.010 0.010 0.042 0.023
25
5.5727*** 0.256 5.4939*** 0.258 1.1945 Yes Yes
_cons D. C. Dummy Rsquared Prob(F) Root MSE Dep. Var. N. of obs 0.8867 0 0.23267 lnprice 448 0.8879 0 0.23427 lnprice 448
Notes: The models in the table are estimated by the method of OLS with strong standard errors. *, **, *** respectively represent significance levels at 10%, 5% and 1% respectively. Checking VIF variables in the model without multicollinearity signs
(Source: Estimates based on survey data of the study)
The price model results explain 89% and the time on market model
explains 30%, similar to the studies cited. The coefficients do not
change between the estimated models; it implies the adding of
Oldflood variable is stability. The models 13 and 15 show that the
flooded of the prior house has a significant effect on the willingnes
to pay of homebuyer. However, the effect on time on market is
negative (line with author's theoritical model) but insignificance. The
listing time is may not a good representation of the buyer's search
time.
This practical test result has contributed to advocating conclusions
from the theoretical model of the effect of homebuyers’ old houses
on their current purchase in section 4.1.
Chapter 5: Conclusions and recommendations
5.1 Conclusion on the results of the thesis
5.1.1 Conclusions on measuring the impact of pricing strategies
26
From the theoretical analysis of the correlation between the listing
price, the selling price and the time on market, Cheng & et al (2008),
Taylor (1999), Sun & Seiler (2013), two hypotheses are proposed:
H0: A overlisting price strategy will signal a high selling threshold and thus increase the selling price of a house but at the same time
prolong the time on market.
H1: An underlisting price strategy will be a signal to the buyer about the possibility of a problem with the house, and therefore the house
will become harder to sell with a reduced price and a long listing
period.
Based on survey data of 448 housing transactions in many districts
of Ho Chi Minh City, the author has tested the impact of the pricing
strategy on the selling price, forsale time, and saleability based on a
3step method: Step 1: the author has established a hedonic model of
house prices from the properties of the house, this price becomes the
basis for determining that the seller's bidding strategy (DOP) is the
difference between the actual listing price and the expected market
price. Step 2: the author has measured the impact of this DOP listing
price strategy on the selling price (due to the proximity between
DOP and selling price, the author then replaced this with the dummy
variable Dum_DOP in this case) on Selling time is based on hedonic
model with robust estimation. The results show that the overlisting
price strategy has the effect of increasing the selling price and
shortening the time on market of houses. Therefore, the author drew
the conclusion of rejecting H0 and accepting H1. Step 3: Apply the
Cox model with 4 different time period to measure the impact of the
listing price strategy on the probability of sale of the house and the
fluctuation of this impact over the time of sale, this is one of new
27
points of the thesis because no studies have investigated this issue.
The results show that the overlisting price strategy has the effect of
increasing the probability of sale of houses, and is strongest in the
first 30 days of sale, after that the effect diminishes and disappears
after 180 days of sale.
5.1.2 Conclusions on developing theoretical framework According to the author, the experiences that the buyers have their old houses will influence their current selection of new housing, but there is no theory to analyze this impact, so the author has developed a theoretical framework to analyze the impact of homebuyers’ old houses on their current buying behaviour, and this is one of the new contributions of the thesis. The results develop a theoretical framework that concludes that the experience of home buyers with their old houses has an impact on their search for new housing. In particular, buyers who with old lowbenefit houses are more likely to find a new home because of the low threshold of benefit, they are willing to pay more and are expected to have a shorter search. And the empirical test results from 448 survey transactions also support this conclusion of the theoretical framework.
ễ ị ể ị LIST OF PUBLISHED WORKS ươ ệ ồ 1. Nguy n Th Bích H ng và Tr ng Thành Hi p, 2019. Ki m đ nh
ưở ỳ ị ố ớ ủ ườ ự ả s nh h ng c a tâm lý k th đ i v i hành vi ng i mua nhà:
ị ườ ứ ở ệ ứ Nghiên c u Kinh t ế , Nghiên c u trên th tr ng nhà Vi t Nam.
2(489), trang 43 – 54.
ễ ồ ị ươ ệ 2. Nguy n Th Bích H ng và Tr ế ng Thành Hi p, 2018. Lý thuy t
ủ ế ườ Ả ưở ủ ở tìm ki m c a ng i mua nhà: nh h ng c a căn nhà đang lên
ế ị ủ ườ ạ T p chí Tài chínhMarketing, 46 quy t đ nh c a ng i mua nhà. ,
trang 1 – 5.
ễ ồ ị ị ậ 3. Phan Đình Nguyên, Tô Th Nh t Minh, Nguy n Th Bích H ng,
ự ạ ả ị ườ ể 2018. Th c tr ng và gi i pháp phát tri n th tr ả ấ ộ ng b t đ ng s n
ạ ồ ạ ế ể T p chí Kinh t và Phát tri n, 252 t i TP.H Chí Minh. (8), trang
30 – 38.
4. Nguyen, H.T.B, Nguyen, H.T., Truong, H.T. (2020). The role of
Listing Price Strategies on the Probability of Sale of a House:
Evidence from Vietnam. Real Estate Management and Valuation,
28(2), 63 – 75. DOI: 10.1515/remav20200016.
