TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 12, No. 11, Novembe
r
2014, pp. 79
0
6
~ 791
1
DOI: 10.115
9
1
/telkomni
ka.
v
12i11.65
30
7906
Re
cei
v
ed
Jul
y
28, 201
4; Revi
sed Septe
m
ber
4, 2014
; Accepte
d
Septem
ber 26,
2014
Model for Alliance Par
t
ners Selection Based on the
Grey Model and DEA Application-Ca
se by Vietnamese
Bank Industry
Chia-
Nan
Wang, Van- Th
anh Phan*
Dep
a
rtment of Industria
l
Engi
neer
ing a
nd M
ana
geme
n
t,
Natio
nal Ka
ohs
iun
g
Univ
ersit
y
of Applie
d Sci
ences, T
a
i
w
a
n
*Corres
pon
di
n
g
author em
ail:
thanhkem
27
1
0
@gma
il.com
A
b
st
r
a
ct
F
a
cing t
he
op
eratio
nal
in
efficiency s
i
tuati
o
n d
u
rin
g
a
lo
n
g
ti
me, th
e sh
arply
bl
oo
mi
ng
of s
m
al
l
do
mestic
ally
b
anks
and
the
b
anki
ng
ba
d d
e
b
t situati
on
inc
r
ease
year
by
year, a
ll
ele
m
e
n
ts hav
e re
duc
e
d
their c
o
mpetiv
eness. T
h
eref
ore, h
o
w
en
ha
ncin
g the
i
r
c
o
m
p
etitiveness? This study based on the
Grey
forecastin
g
mo
deli
ng (GM) a
nd D
a
ta Env
e
l
o
p
m
e
n
t Ana
l
ys
is (DEA) as fo
und
atio
n pro
p
o
ses a
n
effecti
v
e
appr
oach
for h
e
lpi
ng
man
a
g
e
r
find
out the
b
e
st partn
er
w
hen for
m
e
d
a
lli
a
n
ce. Re
alistic
data
of 21
ba
n
ks
w
e
re collecte
d
from the Vietn
a
m sto
ck exch
ang
e, the state bank
of Vietn
a
m a
nd the
i
r official w
ebsite, t
h
e
empiric
a
l study
indicat
e
s that there
are 6 the
best comb
in
ations in the
tota
l of 210 virtual
alli
anc
es. T
hes
e
results ar
e g
o
o
d
sou
n
d
for h
e
l
p
in
g or
gan
i
z
a
t
i
on to s
e
l
e
ct the best c
a
n
d
id
a
t
es w
hen i
m
pl
e
m
e
n
tin
g
a
lli
anc
e.
T
h
is issue ca
n be exten
d
e
d
a
nd ap
pli
ed i
n
many fiel
ds by
consi
deri
ng l
o
ts of different factors in the futur
e
.
Ke
y
w
ords
:
ba
nks, alli
ance st
rategy, Grey forecast
in
g mod
e
lin
g, data e
n
v
e
lo
p
m
ent a
naly
s
is
Copy
right
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Like othe
r co
untrie
s
, ban
king system i
s
lif
eblood of
Vietnam’s e
c
on
omy. It plays an
importa
nt rol
e
in
the
eco
nomy
stabilit
y and
dev
el
opment
of
country. Espe
cially, it i
s
v
e
r
y
importa
nt for developin
g
countrie
s
like Vietnam. Ba
n
ks pl
ays a vital role to re
p
e
l and control
the
inflation, ste
p
by step keep sta
b
le
of m
oney value an
d rat
e
of excha
n
ge, and imp
r
ove
macroe
co
no
mic, inve
stment and
bu
sine
ss envi
r
onment. It
also
co
ntrib
u
tes to
pro
m
ote
investment, p
r
odu
ction a
n
d
export activities.
In the re
ce
ntly years, e
s
peci
a
lly sin
c
e Vi
etnam h
ad no
rmali
z
a
t
ion relatio
n
ship wit
h
orga
nization
s as:
Wo
rld B
ank (WB), In
ternatio
n
a
l M
onetary
Fund
(IMF), A
s
ian
Develo
pme
n
t
Bank (A
DB), and World T
r
ade Organi
za
tion (WT
O),
ban
king
syst
em re-affirme
d
critical role
as
well a
s
po
sition in the su
pport a
nd promotion of
e
c
on
omy. Accordin
g to the
state ban
k
of
Vietnam [1], each year
b
anki
ng
syste
m
ha
s c
ontri
buted ove
r
1
0
% in the o
v
erall e
c
on
o
m
ic
gro
w
th of th
e
co
untry, sol
v
ing and
ge
n
e
rating
job
fo
r thou
sa
nd
workers,
spe
n
d
ing tho
u
san
d
s
billion VND
capital
credit
s
invests
wi
th the dev
el
opment of
economic- social
infrastruct
u
re,
agri
c
ultu
ral le
nding,
rurality, export, smal
l and
mediu
m
-si
z
ed
- e
n
terprises. B
a
n
k
ing
servi
c
e
s
a
r
e
not only limited in the scope of rai
s
in
g capital
and
granta
b
le credit, but also
many kind
s of
mode
rn se
rvice
s
wa
s appl
ied and b
e
ca
me popul
ar a
s
debit card
s,
banki
ng ele
c
tronic
se
rvice
s
,
forex tradi
ng.
Bankin
g net
work
wa
s exp
ande
d a
c
ro
ss the co
untry.
This
ha
s en
a
b
ling
conve
n
i
e
n
t
for peopl
e an
d busi
n
e
s
ses
with ea
sy access to ban
kin
g
servi
c
e
s
Beside
s the achi
evement
s of banki
ng system
ha
s g
a
ined, it also facing ma
ny probl
em
s
and reve
aling
some
sho
r
tcoming
s
. More
spe
c
ificatio
n as follo
ws:
Firstly is the majority of co
mmercial ba
n
ks o
perationa
l inefficiency
for a long time. The
r
e
se
ar
ch
’s res
u
lt in th
e re
fe
r
e
nc
e [2
] indic
a
te
d th
a
t
amo
n
g
3
1
c
o
mme
rc
ia
l b
ank
s
h
a
v
e
c
h
os
e
n
to analyze;
80% re
sp
o
nded i
nefficie
n
t in ban
kin
g
activities, b
u
t just o
n
ly 19% ab
solut
e
ly
efficien
cy an
d nea
rly effici
ency. The
s
e i
s
sue
s
indi
cat
e
that the op
eration
a
l ineff
i
cien
cy for ye
ars
of comme
rcia
l banks.
Secon
d
ly is t
he
sha
r
ply bl
ooming
of
small dom
esti
cally ba
nks i
n
the
sho
r
t time pe
riod.
Acco
rdi
ng to
the state
ban
k of
Vietna
m
at the en
d of
2012, the
sy
stem incl
ude
s
5 state
-
o
w
ne
d
comm
ercial
ban
ks, o
ne
ban
k for
so
cial p
o
licie
s,
one d
e
velo
pment ba
nk,
35 Joint-
stock
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Model for Alli
ance Partne
rs Selectio
n Base
d on the
Gre
y
Mod
e
l a
nd DEA… (Chia- Nan Wan
g
)
7907
comm
ercial
ban
ks,
48
branche
s of
fo
reign
ba
nks,
5 j
o
int- ve
n
t
ure
ban
ks,
5 wholly fo
re
ign-
own
ed
ban
ks, 49
rep
r
e
s
en
tative foreign
ban
ks, 18
finan
ce
com
p
a
n
ies,
12 l
e
a
s
i
ng
comp
anie
s
,
one central p
eople’
s credit
fund with more than 2
4
bran
che
s
spre
ad of count
ry, and more th
an
1000
cre
d
it funds. Althoug
h the sy
stem
appe
ars more and mo
re b
anks with vari
ety forms abo
ut
ownership
and business
types but
m
o
st of them
still have a
sm
all
-
scal
e. Accordi
ng to the
referen
c
e [3]
,
the autho
ri
zed
ca
pital
of larg
er
co
mmercial b
a
n
ks in Vietn
a
m like Agri
ban
k,
Vietcomb
an
k, Vietinbank o
r
BIDV is too small;
just ha
ve nearly 800
million USD l
e
ss much tha
n
small a
nd m
edium
-si
z
e
d
comm
ercial b
anks in
the
same area
s. For in
stants
DBS Bank (9,6
23
million USD);
Unite
d
overseas
Ban
k
(6,
297 millio
n
USD) i
n
Sin
g
a
pore,
Mayba
n
k
(4,1
02
million
USD) in Malaysia, Bang
kok Ban
k
(3,1
78 million US
D) in Thail
a
n
d
Bank Man
d
i
ri (2,122 milli
on
USD); Bank
BNI (1,499 mi
llion USD) in Indon
esi
a
.
Finally i
s
the
ban
kin
g
b
a
d
de
bt
situat
ion in
crea
se
s yea
r
by y
ear. A
c
cordi
ng to
the
monitori
ng re
sults
of state ban
k
[1],
the
bad debt
of
system in
crea
sed
ne
arly
5 t
i
mes in th
e
short
time, Espe
cia
lly, in the mid of 201
2, the bad
debt o
f
system i
s
2
02.099 th
ou
sand billio
n V
ND
“incre
ased 2.5 times in 6 months
com
p
are with ye
a
r
of 2011”, a
c
counting for 8.
6% total loans in
whi
c
h the b
ad debt of state- o
w
ne
d
comme
rci
a
l
banks is 1
25.8 thousa
nd billion VND,
accou
n
ting fo
r 10.37 % total loan
s of st
ate-
o
w
ned
b
anks g
r
oup.
The ba
d debt
of comme
rci
a
l
ban
ks i
s
60.9
thousa
nd billi
on VND, a
c
counting
for 5.
8% total loans of their grou
p.
All above issue
s
an
d sh
ortco
m
ing
s
h
a
ve
red
u
ced
their co
mp
etitiveness. So, how
enha
nci
ng th
eir competiti
v
eness? T
h
i
s
pap
er
p
r
o
v
ides an eff
e
ctive app
ro
ach fo
r help
i
ng
manag
er to fi
nd o
u
t the
be
st pa
rtne
r
wh
en fo
rmed
alli
ance b
a
sed
o
n
the
GM
and
DEA. Fi
rst,
we
use
the
grey
model
predi
cts the
inp
u
t a
nd o
u
tput
fa
ctors in the
fut
u
re
rely
on th
e previou
s
d
a
t
a
aim to
kno
w
t
he pe
rforman
c
e of
DM
Us i
n
the
futu
re.
After that, usi
ng data
envel
opment
analy
s
is
and h
e
u
r
isti
c tech
nique
e
v
aluates
ope
rational
effici
ency
before
and afte
r formed a
n
allia
nce.
Then, we b
a
sed on the cha
r
acte
ri
stic of
DEA to find out the best pa
rtners.
The re
st of this pap
er is
orga
nized a
s
follows. Sect
ion 2 provide
s
pro
p
o
s
ed a
ppro
a
ch.
The research
result
s were
show in
sect
ion 3.
The last se
ction illustrates
so
m
e
concl
u
si
on
and
s
u
gges
tion of
res
e
arc
h
in t
he future.
2.
Proposed Appro
ach
es
Reali
z
e
d
the power a
s
well
as usa
b
ility of GM
and DEA in the real case [3, 4], thi
s
pape
r
combi
ned th
e GM an
d t
he DEA p
r
o
poses
a ne
w system
atic
approa
ch to
find out the
best
allian
c
e mem
bers when im
plementin
g st
rategy. Fo
r fi
nding p
a
rtne
r pro
c
e
ssi
ng i
s
implem
enti
n
g
throug
h eig
h
t step
s. They
are
DM
Us
co
llection,
Input
/ output varia
b
les
sele
ctio
n, Input/ outp
u
t
variable
s
foreca
st by
Grey predi
ction
,
erro
r che
c
king,
Co
rrela
t
ion an
alysi
s
, DEA mo
d
e
l
choosi
ng, Analysis before
Strategi
c Alli
ance, Analysi
s
virtual DM
Us
after Strategic Alliance, and
Partner Sel
e
ction whi
c
h sh
ow in Figu
re
1.
Figure 1. Pro
ducer App
r
oa
ch
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 11, Novem
ber 20
14: 79
06 – 791
1
7908
Explanation
s
of Figure 1 a
r
e sho
w
n b
e
lo
w.
Step1: DMUs Selection
Searchin
g en
tire ban
ks in
Vietnam to find all potential
candi
date
s
to be our
DM
U list
colle
ct histo
r
y data on ca
nd
idate’s b
a
n
ks
Step2: Input and Outp
ut Variabl
es Sele
ction
Rea
d
som
e
p
r
eviou
s
pap
ers and li
sts po
pular va
riable
s
in this field.
Use co
rrelati
on analy
s
is to
analyze p
o
sit
i
ve re
lation
shi
p
betwe
en ea
ch varia
b
le of
inputs
and outp
u
ts.
Verify positive relation
shi
p
betwee
n
ea
ch
variable n
e
gative coeffici
ent betwe
en i
nputs
and outp
u
ts, they need to b
e
remove
d for followin
g
DEA’s basi
c
a
s
sumptio
n
Step 3: Variables Fo
re
ca
st
by Grey Pre
d
iction
Cho
o
se the g
r
ey pre
d
ictio
n
model
Based o
n
the
previou
s
dat
a durin
g 200
8
-2011 to fore
ca
sts the inp
u
ts and o
u
tpu
t
s
Step 4: Accu
racy Ch
eckin
g
:
List co
mmon
indexe
s
to measure t
he a
c
curacy of fore
ca
st model
Step 5: Pearson Co
rrelatio
n
As above me
ntion, Pearso
n correl
ation i
n
D
EA is an i
ndex to test the relatio
n
shi
p
betwe
en inp
u
t
s and outp
u
ts.
Step 6: DEA Model Choo
si
ng
Based o
n
the
characte
ri
stic of each DE
A model to ch
oose model
Step7: Evalu
a
te perfo
rm
ance b
e
fore St
rategi
c Allian
c
e.
We u
s
ed
DEA-Solver Pro
softwa
r
e to
g
e
t the performance of all DMU
s
before
Alliance.
Step8: Firm
ed and Eval
uat
e perfo
rm
ance after Strate
gic Allian
c
e.
The first, we
combi
ne each bank with the re
st of ones to be many
virtual alliance.
And then we
use
DEA-Sol
v
er Pro software run ag
ain
of total DMU
s
Based o
n
the
rank a
nd the
score of
DEA to group
pot
ential efficien
cy of the virtual allian
c
es
Step9: Partne
r Selectio
n
Based o
n
the
rank a
nd the
score of virtu
a
l allian
c
es, t
he analy
s
e
s
of empiri
cal result
s
split into thre
e grou
ps a
nd
interp
ret as b
e
low:
Grou
p 1:
The
banks who g
e
t better re
sul
t
after strategi
c allian
c
e a
n
d
also ma
ke
their pa
rtnership more efficient are the fi
rst prio
rity can
d
idate. The
s
e
candi
date
s
have the goo
d cha
r
a
c
teri
st
ic and n
e
cessarily match
wi
th candi
date
s
’ desire in doi
ng
busi
n
e
ss.
Grou
p 2:
Whi
c
h DM
U in
cre
a
sin
g
perfo
rm
ance after strategic alli
ances while othe
r
DMU
will get a worst performanc
e is the
second priorit
y
.
Grou
p 3:
DM
Us
whi
c
h get
worst or n
o
a
n
y improvem
ent after strat
egic alli
ances
are not reco
mmend
ed in this re
se
arch.
No ne
ed to p
u
t in any effort for alliance
becau
se no a
n
y benefits fo
r both candid
a
tes an
d targ
et candi
date
s
.
3.
Rese
arch Resul
t
s
After su
rvey f
r
om
entire
ba
nkin
g
system
in
Vietna
m,
21 b
a
n
k
s wit
h
complete
d
a
ta are
sele
cted to
b
e
our DM
Us. All informatio
n is
co
lle
cted
from Vietna
m sto
ck
exch
ange,
web
s
it
e of
the state
b
a
n
k
of
Vietnam
[1, 2] an
d [5,
6]. For
data
set is pri
m
arily
drawn from
annu
al fina
ncial
repo
rts o
n
their official we
bsite du
ring 2
008 - 20
11.
Input and out
put variable
s
sele
ction p
r
o
c
e
ss
wa
s sel
e
cted
ca
reful
ness. Based
on so
me
previou
s
p
a
p
e
rs such a
s
[7, 8], we li
ste
d
co
mmon
variable
s
wa
s u
s
ed i
n
this fie
l
d and
after t
hat
we al
so use the Pearso
n correl
ation by DEA “se
e
Pe
arson Correla
t
ion” to che
ck again then we
cho
o
se in
put
and
output
s v
a
riabl
es.
He
n
c
e, o
u
tputs in
this pa
per wil
l
then
in
clude
total lo
an
s (T
L
inclu
d
e
s
total
cu
stome
r
lo
ans
and
total
s
othe
r
le
ndi
ng) and
th
e net
profit
s (NP, the amou
nt of
incom
e
mon
e
y
earning aft
e
r tax). The
main input
s will then in
clu
de total depo
sits (T
D, incl
ude
s
depo
sits fro
m
cu
stome
r
s and othe
r ban
ks), fix
ed assets (FA, comp
osed
of land, pro
p
e
rty
equipm
ents…etc) a
nd t
he o
perating
expen
se
s
(O
E, perso
nn
el expe
nses,
dep
re
ciation
and
arom
atizatio
n
charge
s and
ot
her op
erating expen
se
s).
After finishe
d
DMUs
colle
ction a
nd va
riable
s
settin
g
pro
c
e
s
s, al
l origin
al dat
a of 21
ban
ks in 2
0
11 wa
s sho
w
n in Tabl
e
1. Becau
s
e
of the convenient of
Grey predi
ction
only
requi
re
s a small numb
e
r
of data to proce
s
s a pro
b
l
em (at lea
s
t 4 previou
s
times) and e
a
sy
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Model for Alli
ance Partne
rs Selectio
n Base
d on the
Gre
y
Mod
e
l a
nd DEA… (Chia- Nan Wan
g
)
7909
usin
g. Mo
reo
v
er, the
histo
r
ical
data
of some b
a
n
k
s
in
Vietnam are
incom
p
lete.
Therefore, Grey
model is
suit
able to estab
lish the fore
cast
model. Base
d on the previou
s
dat
a durin
g 200
8-
2011, we use
Grey pre
d
icti
on fore
ca
sts t
he
input an
d output variabl
es in the future.
Table1. O
r
igi
nal Data of All DMU
s
in
2
011
DMUs
Inpu
ts (Billio
n V
ND)
Ou
tpu
t
s (Billio
n
VND)
TD
F
A
OE
TL
NP
DMU
1
176,932
1,237
3,147
184,093
3,208
DMU
2
41,799
1,224
1,296
47,555
947
DMU
3
56,110
232
615
43,173
488
DMU
4
29,810
715
866
29,810
314
DMU
5
136,781
1,191
2,099
105,753
4,114
DMU
6
12,001
827
248
10,009
241
DMU
7
18,298
140
394
15,814
166
DMU
8
55,000
371
1,302
55,000
800
DMU
9
30,774
328
595
22,837
426
DMU
10
30,310
241
516
30,310
234
DMU
11
16,484
381
424
16,825
303
DMU
12
116,221
1,551
1,881
99,619
1,915
DMU
13
72,846
293
1,696
71,475
639
DMU
14
87,916
3,708
3,589
90,161
1,996
DMU
15
12,571
340
331
13,451
248
DMU
16
274,979
2,606
5,700
309,095
4,217
DMU
17
14,283
267
453
13,332
446
DMU
18
125,512
1,913
1,910
138,574
3,039
DMU
19
331,682
3,746
9,078
355,850
6,244
DMU
20
36,356
1,089
938
34,408
84
DMU
21
16,661
345
270
20,616
408
The che
cki
ng
error result o
f
foreca
st mo
del is
ve
ry important ai
m to kno
w
in
whi
c
h the fit
forecast
mod
e
l. No
wad
a
ys, there a
r
e l
o
ts of
in
dex
to mea
s
ure t
he e
rro
r fo
re
ca
st like:
Me
an
squ
a
re
d e
r
ro
r (MSE), Me
an a
b
solute
deviation
(MAD), M
ean
s a
b
solute
p
e
rcentag
e e
r
ror
(MAPE). Thi
s
study uses the MAPE
(Means
absolute
percentage error) to evaluate th
e
acc
u
rac
y
of forec
a
s
t. All res
u
lt was
showed as
follows
:
Table 2. Average MAPE Error of DMU
s
DMUs
A
v
e
r
age
M
A
PE
DMU
1
2.74%
DMU
2
1.89%
DMU
3
9.23%
DMU
4
6.01%
DMU
5
4.07%
DMU
6
4.00%
DMU
7
1.86%
DMU
8
2.90%
DMU
9
3.23%
DMU
10
4.45%
DMU
11
3.63%
DMU
12
4.14%
DMU
13
5.94%
DMU
14
5.10%
DMU
15
5.41%
DMU
16
3.06%
DMU
17
5.97%
DMU
18
7.09%
DMU
19
5.35%
DMU
20
3.87%
DMU
21
5.05%
Average MAPE o
f
21 DMU
s
4.52%
This table ind
i
cated that the error value
of
foreca
st is
very low less than 10% [9], which
confirm that
GM mo
del i
s
fit model in
this
ca
se
st
udy. The
r
efo
r
e, this mea
n
s the
result
s of
forecast in ta
ble 4 have a
high reli
abilit
y. The supe
r- SBM model
has mo
re a
d
vantage
s th
an
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 11, Novem
ber 20
14: 79
06 – 791
1
7910
traditional
DE
A model [11-13], Tone et
al. [11,
13] show that
sup
e
r- SBM mo
del co
uld ran
k
s
extreme
DEA efficient DM
Us, i
s
able to
eliminat
e the
dra
w
ba
ck of
earlie
r mo
de
l. Furthermore,
the ch
ara
c
te
ristic of
comm
erci
al ba
nks i
n
Vi
etnam a
r
e small, the
quite differe
n
t
among b
a
n
ks
related to aut
hori
z
ed
capit
a
l. Therefo
r
e,
supe
r-
SBM
model is
suita
b
le in this case study.
As above m
ention, Pearson co
rrelatio
n in DEA is an index to
test the rel
a
tionship
betwe
en in
pu
ts and
outp
u
ts. The
re
sult
s sho
w
t
hat
correl
ation
coefficient b
e
twee
n inp
u
t a
nd
output va
riab
les
are hi
ghl
y positive
“m
ore
than
0.
7”,
whi
c
h exhib
i
ts
go
od co
rrelation and
well
compli
es
with
the prerequi
site conditio
n
of the DEA model.
For the alli
an
ce impl
ement
ation, the first
,
we
co
mbin
e
each ban
k
wi
th the re
st of one
s to
be m
any virt
ual alli
an
ce
s. The
totals h
a
ve 21
0 virtu
a
l allia
nces.
Then,
we
u
s
e the
supe
r-SBM
model to me
asu
r
e the efficien
cy of all the virtual
ba
nks. Acco
rdi
ng to the result, in total 2
10
virtual alliances, there are
6 the best combinat
ions. T
hese were show in bel
ow t
able.
Table 3. The
First Priority Group for
Strategic Alliance Based on the Rank
Virtual ran
k
Virtual allia
nce
Virtual sc
ore all
i
ance
Grou
p
4 DMU
5
+DMU
18
1.177427
1
9 DMU
16
+DMU
18
1.087724
1
10 DMU
1
+DMU
19
1.081508
1
11 DMU
1
+DMU
18
1.079189
1
12 DMU
1
+DMU
16
1.076974
1
13 DMU
1
+DMU
5
1.071406
1
For this g
r
o
u
p
, alliance strategy not onl
y hel
ps target
DMU imp
r
ov
e the perfo
rm
ance but
also
ma
ke th
eir p
a
rtn
e
rshi
p imp
r
ove
pe
rforma
nce.
T
h
is m
ean
s th
ey wo
uld
hav
e st
rong
de
si
re to
form allian
c
e.
So, for this group thi
s
stud
y strongly re
commen
d
s to
corpo
r
ate.
The rest
of combinatio
ns
belon
g to th
e group
2 a
nd the g
r
o
u
p
3, re
spe
c
tively. With
these
combi
n
ations b
e
long
in these gro
ups, th
is
stud
y sugge
sts th
at no need p
u
t in any effort
for alliance because when carry
out alliance strategy,
these combi
nations could make for them
or their p
a
rtn
e
rs o
r
both of
them meet a risk.
4. Conclusio
n
Based
on
fou
ndation t
heo
ry of DEA an
d
GM, this stu
d
y pro
p
o
s
ed
an effe
ctive a
ppro
a
ch
for g
u
iding
m
ange
rs to fin
d
out th
e b
e
st
partne
r
s
whe
n
formi
ng
stra
tegic
allian
c
e
aim to
enh
an
ce
their comp
etitiveness in t
he future. E
s
pe
cially, it will be
com
e
the gold
en key guide fo
r top
policyma
k
e
r
i
n
whi
c
h solving re
stru
cturi
ng economy i
n
Vietnam.
By ours p
r
op
ose
d
effe
ctive ap
pro
a
ch, this
re
sea
r
ch f
ound
that h
a
ve 6
com
b
inati
ons are
the be
st allie
d in the total
of 210 virtu
a
l a
llian
c
e
s
.
These results a
r
e g
ood
sound fo
r h
e
l
p
ing
organization to select the best ca
ndidates when implementing alliance.
The accom
p
l
i
shme
nt
of
t
h
is study ca
n
lea
d
to
future
re
se
arch
with
more i
nput a
n
d
output variabl
es, more wa
ys to combin
e together
in
alliance
wa
s analysis
su
ch as co
mbini
ng
three or fou
r
DMU togeth
e
r, and more
different
industrie
s can b
e
assesse
d
by this propo
sed
approa
ch. F
u
rtherm
o
re, dif
f
erent fo
re
ca
sting
metho
d
and
DEA m
odel
s
can
be
used
to expl
ore
and devel
op i
m
porta
nt issu
es.
Referen
ces
[1]
T
he State bank of Vietnam
w
e
bsite:
[Online] avail
a
b
l
e:
http://
w
w
w
.
s
b
v.gov.
v
n
[2]
HX Ng
u
y
e
n
. Applic
atio
n of D
EA mode
l to
e
v
alu
a
te
the op
eratio
n
effici
en
c
y
of
commerc
ial banks
i
n
Vietnam.
Bank
i
ng revi
ew
. 201
2; 20.
[3]
CN W
a
ng, KZ
Li, W
P
T
s
eng,
KY Li, MY
Ka
n
,
KT
T
s
ai, PH
T
s
ai.
A Stra
te
gi
c Al
li
an
ce Ap
pro
a
c
h fo
r the
Industry of Ra
dio F
r
e
que
n
cy Identificati
on in
T
a
iw
an
. T
e
chnol
og
y Ma
na
g
e
ment for th
e
Globa
l F
u
ture,
PICMET
2006;
1.
[4]
CN W
a
n
g
, VT
Phan. A
lli
anc
e
Partner S
e
l
e
cti
on M
ode
l B
a
se
d o
n
Gre
y
t
h
e
o
r
y
a
nd
Data
Envel
opme
n
t
Anal
ys
is- Case
b
y
b
anki
ng i
n
d
u
str
y
in Vi
etna
m.
Pensee Jo
u
r
nal
. 20
14; 76:
167-
179.
[5]
T
he
w
ebs
ite of the bank
er: [Onlin
e] avai
la
ble
:
http://
w
w
w
.
t
heb
ank
er.com/top10
00.
[6]
T
he Vietnam
stock e
x
ch
an
g
e
w
e
bsite:
[Onlin
e]
av
ai
lab
l
e: http://
w
w
w
.
vietnam-r
eport.
com/vietnam-
stock-ex
c
h
ange
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Model for Alli
ance Partne
rs Selectio
n Base
d on the
Gre
y
Mod
e
l a
nd DEA… (Chia- Nan Wan
g
)
7911
[7]
Mlima AP, Hja
lmarsson L.
M
easur
ement of
Inputs and Out
puts in the
bank
ing i
n
d
u
str
y
.
Ta
n
z
ane
t
Journ
a
l
. 20
02; 3: 12-22.
[8]
Avikiran
NK. A
ssociati
on
of D
EA sup
e
r-effici
enc
y esti
mates w
i
t
h
financ
ial ratios: Invest
i
g
a
t
ing th
e cas
e
for Chin
ese b
a
n
k.
Om
ega
. 20
11; 39: 32
3-33
4.
[9]
Le
w
i
s C
D
. Indu
strial an
d Busi
ness forecasti
n
g method, Butt
er
w
o
rth-H
e
i
n
e
m
ann, Lo
nd
on.
1982.
[10]
T
one K. A sl
ack- b
a
sed
m
easur
e of
effi
cienc
y
in
d
a
ta
env
elo
p
me
nt an
al
ysis.
Res
earch
Re
port.
Graduate Sc
ho
ol of Polic
y Sci
enc
e, Saitam
a
Univers
i
t
y
. 1
9
9
7
.
[11]
K.
T
one: A sla
ck- base
d
me
asure
of
efficie
n
c
y
in d
a
ta e
n
v
elo
p
ment a
n
a
l
y
s
is,
Europ
e
a
n
Jour
nal
of
Operatio
nal R
e
search
, 20
01;
19: 498-
50
9.
[12]
Du J, Liang L
,
and Z
hu J.
A slack-bas
ed
m
easure of e
fficienc
y
in dat
a enve
l
opm
ent
anal
ysis:
a
comment,
Euro
pea
n Jour
nal o
f
Operationa
l R
e
searc
h
. 201
0; 204: 69
4-69
7.
[13]
T
one K. A slac
k-base
d
me
as
ure of s
u
p
e
r-ef
f
icienc
y in data
enve
l
o
p
ment ana
l
y
sis,
Eur
o
pea
n Jo
urn
a
l
of Operation
a
l
Rese
arch
. 20
0
2
: 32-41.
Evaluation Warning : The document was created with Spire.PDF for Python.