# RECORD 154f55871a6811fb1dd94f5c020e97dabe79c8c3
# TITLE: Efficient property value estimation for single-family homes in central Florida
# YEAR: 2024.0  VENUE: International Transactions in Operational Research  DOI: 10.1111/itor.13565
# AUTHORS: S. Lozano; E. Gutiérrez; Kotryna Klizentyte; A. Susaeta
# PAGES: 29  CANDIDATE TABLE PAGES (1-based): [10, 16, 17, 18, 19]
# ABSTRACT: This paper proposes the use of a metafrontier efficiency analysis approach to estimate residential property values. The objective is to mitigate inefficiencies in these valuations, thereby enhancing tax revenue and fairness in tax burdens. Property features (such as acreage, living area, number of bedrooms, number of bathrooms, age, and pool presence) along with contextual variables (such as county and median household income) are utilized to estimate just market value. Observations can be grouped based on some of these variables, allowing a metafrontier analysis to be conducted. This analysis not only assesses efficiency scores but also identifies technology gaps for each group. Furthermore, the proposed approach can be employed to estimate the value of any property based on its characteristics. To illustrate the proposed approach, we utilize data on over 30,000 properties in Orange and Seminole Counties in central Florida. Additionally, we present the results of a robust regression approach for comparison purposes. While our approach can serve as a valuable tool across various sectors, its primary aim is to promote best practices in governance regarding land use and real estate valuation. This includes balancing multiple objectives such as tax revenue, economic growth, environmental sustainability, and social considerations.


===== PAGE 1 (score 20, numeric-paren 0, table-words 1, var-words 18) =====

Intl. Trans. in Op. Res. 0 (2024) 1–30
DOI: 10.1111/itor.13565
INTERNATIONAL
TRANSACTIONS
IN OPERATIONAL
RESEARCH
Eﬃcient property value estimation for single-family homes in
central Florida
Sebastián Lozanoa
, Ester Gutiérreza,∗
, Kotryna Klizentyteb
and Andrés Susaetac
aDepartamento de Organización Industrial y Gestión de Empresas I, Universidad de Sevilla, Sevilla, Spain
bNatural Resource Economics and Policy, School of Forest, Fisheries, and Geomatics Sciences, University of Florida,
Gainesville, FL, USA
cDepartment of Forest Engineering, Resources and Management, Oregon State University, Corvallis, OR, USA
E-mail: slozano@us.es[Lozano]; egm@us.es [Gutiérrez]; kklizentyte@uﬂ.edu[Klizentyte];
andres.susaeta@oregonstate.edu[Susaeta]
Received 20 February 2024; received in revised form 17 July 2024; accepted 30 September 2024
Abstract
This paper proposes the use of a metafrontier eﬃciency analysis approach to estimate residential property val-
ues. The objective is to mitigate ineﬃciencies in these valuations, thereby enhancing tax revenue and fairness
in tax burdens. Property features (such as acreage, living area, number of bedrooms, number of bathrooms,
age, and pool presence) along with contextual variables (such as county and median household income)
are utilized to estimate just market value. Observations can be grouped based on some of these variables,
allowing a metafrontier analysis to be conducted. This analysis not only assesses eﬃciency scores but also
identiﬁes technology gaps for each group. Furthermore, the proposed approach can be employed to estimate
the value of any property based on its characteristics. To illustrate the proposed approach, we utilize data
on over 30,000 properties in Orange and Seminole Counties in central Florida. Additionally, we present the
results of a robust regression approach for comparison purposes. While our approach can serve as a valuable
tool across various sectors, its primary aim is to promote best practices in governance regarding land use
and real estate valuation. This includes balancing multiple objectives such as tax revenue, economic growth,
environmental sustainability, and social considerations.
Keywords: residential property value; local tax revenue; tax burden fairness; non-parametric metafrontier eﬃciency anal-
ysis; Free Disposal Hull (FDH)
1. Introduction
Housing prices, representing the value at which a property is sold in the open market, serve as
crucial economic indicators (Gabrielli and French, 2021). These prices not only reﬂect the overall
∗Corresponding author.
© 2024 The Author(s).
International Transactions in Operational Research published by John Wiley & Sons Ltd on behalf of International Federation
of Operational Research Societies.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which
permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no
modiﬁcations or adaptations are made.


===== PAGE 10 (score 25, numeric-paren 0, table-words 3, var-words 16) =====

10
S. Lozano et al. / Intl. Trans. in Op. Res. 0 (2024) 1–30
Table 1
Descriptive statistics for the clean dataset (n = 30,555).
Min.
Max
25th
percentile
Median
Mean
75th
percentile
IQR
Std. dev
JMV
58,373
1,477,344
217,944
276,355
312,516
361,977
144,033
149,651.8
LIVARE
672
9298
1757
2224
2394
2831
1074
931.13
AGE (yrs)
4
121
21
34
33.01
43
22
13.43
ACREAGE
0.020
5.990
0.210
0.288
0.402
0.426
0.216
0.401
MHHI
51,386
97,185
56,209
79,252
75,943
95,239
39,030
15,644
NBEDS
1.00
7.00
2.00
3.00
3.00
4.00
2.00
0.92
NBATHS
1.00
6.00
3.00
3.00
3.36
4.00
1.00
0.84
County
n
%
Orange
7043
23%
Seminole
23,512
77%
Pool
n
%
Yes
15,577
51%
No
14,978
49%
Note: IQR (interquartile range) = 75th −25th percentile.
of those observations. Figure 3 illustrates the results, showing the dispersion plot of observed JMV
versus the deviation of observed JMV from estimated JMV values (in percentage). The visual
comparison between Figs. 3a,b reveals the improvement in the residuals of the regression model
after the removal of outliers, with the average absolute deviation decreasing from 11.49% to 9.43%.
3.2. Non-convex deterministic property valuation model
In this section, an FDH property valuation model is presented. Due to the discrete nature of
some variables, a non-convex DEA technology is assumed. The model assumes VRS to account
for scale size eﬀects and considers the JMV of the property as the primary output variable,
while the property features, such as acreage, living area, number of bedrooms, and number of
bathrooms, are modeled as non-discretionary inputs. The median household income of the area
is also considered as a non-discretionary input as is the age of the property. This classiﬁcation
of the non-discretionary variables as input or outputs is based on their expected eﬀect on the
JMV output. Thus, it is assumed that, ceteris paribus, the larger the input variables the larger the
corresponding JMV should be and that, ceteris paribus, the older the property, the lower its JMV
should be. Following the approach outlined by Banker and Morey (1984), a property is deemed
ineﬃciently valued if another property with less acreage, less living area, and lower number of
bedrooms and bathrooms (and less median household income and a greater age) has a higher
JMV. The model computes, for a certain property and given its characteristics, the maximum JMV
it could be assigned based on the JMV of comparable properties in the sample. Furthermore, this
ability to compute the property value using as benchmarks speciﬁc existing observations (and not
a ﬁctitious target computed as a convex combination of observed properties) is one of the strengths
of FDH.
© 2024 The Author(s).
International Transactions in Operational Research published by John Wiley & Sons Ltd on behalf of International Federation
of Operational Research Societies.
 14753995, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/itor.13565 by Readcube (Labtiva Inc.), Wiley Online Library on [17/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License


===== PAGE 11 (score 14, numeric-paren 0, table-words 0, var-words 14) =====

S. Lozano et al. / Intl. Trans. in Op. Res. 0 (2024) 1–30
11
Before formulating the model, let us introduce the notation:
Data
J
Set of properties
ACREAGE j
Acreage of property j ∈J
LIVAREAj
Living area (air-conditioned) of property j ∈J
NBEDS j
Number of bedrooms of property j ∈J
NBATHS j
Number of bathrooms of property j ∈J
MHHIj
Median household income of the area of property j ∈J
AGE j
Age of property j ∈J
JMVj
Just Market Value of property j ∈J
0
Index of property whose JMV eﬃciency is being assessed
Decision variables
(λ1, λ2, . . . , λn)
Binary variables used to compute the target JMV of Property 0
TARGETJMV0
Target JMV of Property 0
ξ0
JMV eﬃciency of Property 0
FDH model
Maximize
TARGETJMV0
s.t.

j∈J
λ j · ACREAGE j ≤ACREAGE0

j∈J
λ j · LIVAREAj ≤LIVAREA0

j∈J
λ j · NBEDSj ≤NBEDS0

j∈J
λ j · NBATHSj ≤NBATHS0

j∈J
λ j · MHHIj ≤MHHI0

j∈J
λ j · AGE j ≥AGE0

j∈J
λ j · JMVj ≥TARGETJMV0

j∈J
λ j = 1
λj ∈{0, 1} ∀j ∈J.
(1)
Once the target JMV of DMU 0 is determined, the corresponding JMV eﬃciency score can be
computed as
ξ0 =
JMV0
TARGETJMV ∗
0
.
(2)
Since the solution λ0 = 1, λ j = 0 ∀j ̸= 0, TARGETJMV0 = JMV0 is feasible in (1), it follows
that the optimal solution TARGETJMV ∗
0 ≥JMV0, and, hence, ξ0 ≤1. The JMV undervaluation
© 2024 The Author(s).
International Transactions in Operational Research published by John Wiley & Sons Ltd on behalf of International Federation
of Operational Research Societies.
 14753995, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/itor.13565 by Readcube (Labtiva Inc.), Wiley Online Library on [17/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License


===== PAGE 16 (score 14, numeric-paren 0, table-words 4, var-words 3) =====

16
S. Lozano et al. / Intl. Trans. in Op. Res. 0 (2024) 1–30
Table 2
Average group and metafrontier just market value (JMV) undervaluation for each group and for the whole sample.
County
Pool
No. of
properties
Avg. JMV0
Avg. JMV SLACKg0
0
US$ (also in %)
Avg. JMV SLACKM
0
US$ (also in %)
Seminole
No
10,643
282,799
43,249 (16.2%)
49,116 (18.2%)
Seminole
Yes
12,869
371,208
80,753 (22.4%)
97,620 (27.4%)
Orange
No
4335
223,145
28,803 (12.8%)
83,220 (35.6%)
Orange
Yes
2708
293,456
43,369 (15.0%)
104,303 (34.6%)
ALL
No
14,978
265,534
39,068 (15.2%)
58,986 (23.2%)
ALL
Yes
15,577
357,691
74,254 (21.1%)
98,782 (28.6%)
Seminole
ALL
23,512
331,189
63,776 (19.6%)
75,664 (23.2%)
Orange
ALL
7043
250,180
34,403 (13.6%)
91,326 (35.2%)
ALL
ALL
30,555
312,516
57,006 (18.2%)
79,274 (26.0%)
Table 3
Average technology gap ratio (TGR) and group and metafrontier JMV eﬃciency for each group and for the whole sample.
County
Pool
ξ g
ξ g,M
TGRg
Seminole
No
0.878
0.866
0.982
Seminole
Yes
0.842
0.810
0.964
Orange
No
0.895
0.754
0.822
Orange
Yes
0.878
0.760
0.847
ALL
No
0.883
0.833
0.939
ALL
Yes
0.848
0.801
0.946
Seminole
ALL
0.858
0.835
0.971
Orange
ALL
0.888
0.756
0.833
ALL
ALL
0.865
0.817
0.943
instead of a simple arithmetic average, the single-output case of the TGR aggregation method pro-
posed in Walheer (2018) has been used. As expected, the group eﬃciency scores are always larger
than the corresponding metafrontier eﬃciencies (0.865 vs. 0.817, for the whole sample), and, simi-
larly, the average TGR are consistently less than (or at most equal to) unity. These patterns persist
throughout the table. Consequently, on average, both in the group and metafrontier projections,
properties with pools demonstrate lower JMV eﬃciency, compared to properties without pools
(0.848 vs. 0.883 in one case, and 0.801 vs. 0.833 in the other). However, while the group frontier
projection indicates lower JMV eﬃciency for properties with pools in both counties, the metafron-
tier projection reveals this trend only in Seminole County, not in Orange County. Additionally, it
is evident that the average group eﬃciency of all properties (with and without pools) in Seminole
County is lower than in Orange County, while the opposite holds true for the average metafrontier
eﬃciency.
Regarding TGR, this measures the undervaluation of the best-valued properties within a given
group relative to the best-valued comparable properties across all groups. A TGR value closer to
unity signiﬁes relatively better-valued properties within that group. As per Table 3, the distance
between the group eﬃcient frontier and the metafrontier, aggregated for all properties, yields a
TGR value of 0.943. This value aligns with the aggregate TGR of both with-pool and without-pool
© 2024 The Author(s).
International Transactions in Operational Research published by John Wiley & Sons Ltd on behalf of International Federation
of Operational Research Societies.
 14753995, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/itor.13565 by Readcube (Labtiva Inc.), Wiley Online Library on [17/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License


===== PAGE 17 (score 63, numeric-paren 21, table-words 13, var-words 11) =====

S. Lozano et al. / Intl. Trans. in Op. Res. 0 (2024) 1–30
17
Table 4
Results of the stochastic frontier analysis.
Variables
Coeﬃcient
Standard error
Variables
Coeﬃcient
Standard error
Production frontier
Ineﬃciency
log(LIVAREA)
0.8086***
0.0032
County
0.3097***
(0.0033)
log(NBEDS)
0.0480***
0.0032
Pool
−0.0575***
(0.0032)
log(NBATHS)
−0.0615***
(0.0038)
Constant
0.0262***
(0.0034)
log(MHHI)
0.2214***
(0.0037)
σ u
0.0116**
0.0059
log(ACREAGE)
0.0750***
(0.0013)
σ v
0.1246***
0.0059
log(AGE)
−8***
(0.0024)
λ = σ u/σ v
0.0929***
0.0063
Constant
4.2852***
(0.0440)
Observations
30,555
Wald chi2(6) = 253,536.27***
Log-likelihood = 20,211.15
∗∗Signiﬁcant at 0.05;
∗∗∗Signiﬁcant at 0.01 level
Table 5
Technical eﬃciency (TE) scores.
Technical eﬃciency
Mean
SD
Max
Min
TE
0.9274
0.1084
0.9969
0.7117
TE (Seminole)
0.9862
0.0113
0.9969
0.9686
TE (Orange)
0.7310
0.0206
0.7602
0.7117
TE (with pool)
0.8988
0.1175
0.9786
0.7117
TE (without pool)
0.9549
0.0907
0.9969
0.7540
groups, indicating no signiﬁcant overall diﬀerences between the two groups when within-group un-
dervaluation is factored out. Although the with-pool group marginally outperforms the without-
pool group in Orange County, this is oﬀset by the reverse situation in Seminole County. Nonethe-
less, the impact of County location on property undervaluation appears to outweigh the presence of
pools. Thus, across all properties, including those with and without pools, the best-valued proper-
ties in Seminole County appear to be superior to the best-valued comparable properties in Orange
County.
The results of the SFA estimation are shown in Table 4. All the regressors are signiﬁcant at
the 0.01 level. The signs of the coeﬃcients are as expected, except in the case of the number of
bathrooms, which seems to have a small negative eﬀect on the JMV.
The eﬃciency scores computed by SFA, shown in Table 5, indicate an average value of 0.927
for the whole sample. Contrary to the DEA results, properties without pools show higher eﬃciency
scores than those properties with pools (0.955 vs. 0.898). It can be also seen that the eﬃciency scores
in Seminole County are higher than those in Orange County (0.986 vs. 0.7310), which is consistent
with the TGR results mentioned above.
© 2024 The Author(s).
International Transactions in Operational Research published by John Wiley & Sons Ltd on behalf of International Federation
of Operational Research Societies.
 14753995, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/itor.13565 by Readcube (Labtiva Inc.), Wiley Online Library on [17/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License


===== PAGE 18 (score 53, numeric-paren 20, table-words 11, var-words 7) =====

18
S. Lozano et al. / Intl. Trans. in Op. Res. 0 (2024) 1–30
Table 6
Ordinary least squares coeﬃcient estimates with robust standard estimators and performance accuracy metrics.
Coeﬃcients
Estimate
(std. error)a
CI lower
CI upper
INTERCEPT
4.641****
(0.049)
4.544
4.737
log(LIVAREA)
0.803***
(0.004)
0.796
0.810
log(ACREAGE)
0.078***
(0.001)
0.075
0.080
log(MHHI)
0.222***
(0.004)
0.214
0.230
log(NBEDS)
0.042***
(0.003)
0.034
0.049
log(NBATHS)
−0.060***
(0.004)
−0.067
−0.052
log(AGE)
−0.173***
(0.002)
−0.177
−0.168
County
−0.308***
(0.003)
−0.314
−0.301
Pool
0.024***
(0.002)
0.020
0.028
County*Pool
0.035***
(0.003)
0.029
0.042
Goodness-of-ﬁt statistics
Adjusted R2 = 0.911
AIC = −40553
BIC = −40441
RMSE = 0.124
Note: Dependent variable JMV; number of observations = 30,555.
aHeteroskedasticity-consistent covariance matrix estimator (type H2C).
4.2. Comparison with regression analysis
Before presenting the results of the log–log regression, let us mention that outliers/extreme values
were removed because they increase predictive deviations between the observed and ﬁtted JMV.
Also, to decide the speciﬁcation of the regression, that is, a linear model (linear function of re-
pressors) versus a log-linear model (linear function of logs repressors), the test proposed in Koop
(2008) was implemented. The results conﬁrmed the rejection of the hypothesis that the linear and
log–log model perform equally well. Moreover, since the residual sum of squared (RSS), RSSlinear
= 1036.88>RSSlog-log = 471.75, the log–log model seems preferable to the linear model at 1% sig-
niﬁcance level (lambda statistic = 12,031.5; p < 0.01).
To address the challenges posed by heavy-tailed error distribution, heteroskedasticity in error
variance, and extreme values, a robust estimation procedure was employed. This procedure utilizes
robust standard estimators within the OLS estimation framework (Blair et al., 2022). Additionally,
robust regression via M-estimation, employing the iterative re-weighted least squares estimation
method based on the Huber function, was applied (Venables and Ripley, 2002). Both approaches
yielded similar results. The OLS robust regression coeﬃcients for the log–log model are presented
in Table 6. Note that not only are all the regression coeﬃcients individually highly statistically
© 2024 The Author(s).
International Transactions in Operational Research published by John Wiley & Sons Ltd on behalf of International Federation
of Operational Research Societies.
 14753995, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/itor.13565 by Readcube (Labtiva Inc.), Wiley Online Library on [17/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License


===== PAGE 19 (score 25, numeric-paren 1, table-words 5, var-words 12) =====

S. Lozano et al. / Intl. Trans. in Op. Res. 0 (2024) 1–30
19
signiﬁcant (based on the F-statistic F_value = 2.937e+04; p < 0.01), but it can also be concluded
that jointly the nine variables are highly statistically signiﬁcant. The interpretation of the estimated
coeﬃcients reveals that the most inﬂuential variable aﬀecting the JMV is the LIVAREA so that the
coeﬃcient of Log(LIVAREAi) is 0.803, that is, an increase of 1% in LIVAREA increases the JMV,
on average, ceteris paribus, about 0.8%. The second most prominent variable is determined by the
structural characteristic of the county location so that the coeﬃcient of the County is −0.308, that
is, the average Orange County property prices are 26.5% lower (e−0.308 −1 = −0.265), compared
to Seminole County property prices, if all other variables are kept constant. The presence of a pool
in the property seems to be the least inﬂuential variable, that is, the average JMV increases by just
2.4% (e0.024 −1 = 0.024) if the property has a pool.
Other explanatory variables with an impact on house prices are MHHI (an increase of 1% in
MHHI increases, on average, ceteris paribus, the JMV by 0.22%. Also, as expected, the age of the
property has a negative eﬀect on JMV (−0.173% for an 1% increase in AGE), and the size of the
property has a positive eﬀect on JMV (0.081% for a 1% increase in ACREAGE). On the other hand,
the lower the number of bathrooms and the higher the number of bedrooms, the higher the JMV,
with a slightly higher eﬀect of the NBATHS variable (−0.06% vs. 0.04%). The interaction dummy
variable also has a highly signiﬁcant eﬀect on JMV. Thus, on average, a property located in Orange
County has a lower average JMV (−26.5%), and, on average, the presence of a pool increases JMV
by about 2.4%. And when both circumstances concur, the overall eﬀect on the JMV is a reduction
in the JMV of 22.0% (e−0.308+0.024+0.035 −1 = e0.249 −1 = −0.220). Finally, the model’s accuracy
metrics (Adjusted R2, Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC),
Root Mean Square Error (RMSE)) support its validity.
Note that the results from the SFA estimation (Table 4) show similar outcomes in terms of statis-
tical signiﬁcance and the magnitude of the common coeﬃcients. For example, a 1% increase in the
age of the property reduces the JMV by 0.196%, while a 1% increase in the number of bathrooms
decreases the JMV by 0.0615%. In the SFA model, we use the location of houses and the presence
of pools as factors that explain ineﬃciency, with the results indicating that properties with pools
lead to an increase in the technical eﬃciency of the property appraisal, while properties located in
Seminole County reduce the technical eﬃciency of the property appraisal.
Table 7 shows, for each group pf observations and for the whole sample, the comparison between
the JMV ﬁtted by the regression model (labeled FITTEDJMV reg) and: (a) the observed JMV,
(b) the group frontier target JMV, and (c) the metafrontier target JMV. Note that, since for each
Property 0 it holds that JMV0 ≤TARGETJMV g0
0
≤TARGETJMV M
0 , it follows that these
three ratios are decreasing, that is,
FITTEDJMV reg
0
JMV0
≥
FITTEDJMV reg
0
TARGETJMV
g0
0
≥
FITTEDJMV reg
0
TARGETJMV M
0 . Moreover, since
regression captures the central tendency, that is, it estimates average values instead of best values
as DEA does, the ratios
FITTEDJMV reg
0
TARGETJMV
g0
0
and more so
FITTEDJMV reg
0
TARGETJMV M
0
are below unity, with overall
averages of 0.866 and 0.826, respectively. The average ratio
FITTEDJMV reg
0
JMV0
is, understandably, close
to unity (1.010), with very small diﬀerences between the diﬀerent groups. The relative diﬀerences in
the average value of
FITTEDJMV reg
0
TARGETJMV
g0
0
and
FITTEDJMV reg
0
TARGETJMV M
0
between the diﬀerent groups are a reﬂection
of the corresponding group and metafrontier target JMV mentioned above. For example, since
on average, in both the group and metafrontier projections, those properties with a pool are more
© 2024 The Author(s).
International Transactions in Operational Research published by John Wiley & Sons Ltd on behalf of International Federation
of Operational Research Societies.
 14753995, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/itor.13565 by Readcube (Labtiva Inc.), Wiley Online Library on [17/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License


===== PAGE 20 (score 16, numeric-paren 0, table-words 1, var-words 12) =====

20
S. Lozano et al. / Intl. Trans. in Op. Res. 0 (2024) 1–30
Table 7
Comparison of regression JMV estimates and group and metafrontier targets.
County
Pool
Avg.
FITTEDJMV reg
0
JMV0
Avg.
FITTEDJMV reg
0
TARGETJMV g0
0
Avg.
FITTEDJMV reg
0
TARGETJMV M
0
Seminole
No
1.013
0.881
0.869
Seminole
Yes
1.007
0.840
0.807
Orange
No
1.012
0.898
0.754
Orange
Yes
1.009
0.879
0.758
ALL
No
1.013
0.886
0.835
ALL
Yes
1.007
0.847
0.798
Seminole
ALL
1.010
0.858
0.835
Orange
ALL
1.011
0.891
0.755
ALL
ALL
1.010
0.866
0.816
undervalued than properties without one, both ratios
FITTEDJMV reg
0
TARGETJMV
g0
0
and
FITTEDJMV reg
0
TARGETJMV M
0
are lower for
properties without a pool than for properties with one. Similarly, in the group frontier projection,
properties (including with and without pool) are, on average, more undervalued in Seminole County
than in Orange County, while the opposite happens in the metafrontier projection. Therefore, the
ratio
FITTEDJMV reg
0
TARGETJMV
g0
0
is lower for properties in Seminole County than in Orange County, while, on the
contrary, the ratio
FITTEDJMV reg
0
TARGETJMV M
0 is lower for properties in Orange County than in Seminole County.
4.3. Using the DEA-based property value estimation function
As discussed in Section 3.4, the proposed FDH model serves as an estimation function for the JMV
of a property based on its characteristics and group aﬃliation. Furthermore, it can also estimate
and compare the values of similar properties across diﬀerent groups. To demonstrate the utility
and eﬀectiveness of this capability, we examine how the JMV of a property varies according to
its characteristics, county of location, and the presence or absence of a pool. Consider a property
with median acreage (0.288 acres), median living area (2224 sq ft), median number of bedrooms
(3), median number of bathrooms (3), median MHHI value ($79,252), and median age (between
30 and 40 years). Figures 4–6 depict the corresponding JMV of this property, estimated for the
four groups considered, while varying one property characteristic at a time and keeping the others
at their median values. It is important to note that, apart from the qualitative trends discussed
below, the key strength of using the proposed DEA-based estimation function lies in its ability to
provide quantitative estimates derived from the best-valued similar properties within its group. By
utilizing the estimation function for a diﬀerent group, it is also possible to estimate and compare
the maximum JMV value of a comparable property in that group.
Note, ﬁrst of all, the monotonic, non-concave character of the curves, typical of FDH technology.
In each plot, the metafrontier estimation JMV M is represented by the non-concave curve (not
labeled in the ﬁgures) that envelops the curves of all four groups from above. Regarding the living
area of the property, it is evident that an increase in the living area leads to a higher estimated JMV
© 2024 The Author(s).
International Transactions in Operational Research published by John Wiley & Sons Ltd on behalf of International Federation
of Operational Research Societies.
 14753995, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/itor.13565 by Readcube (Labtiva Inc.), Wiley Online Library on [17/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License