Gia , Yahel; Mi elman, Amichai
A icle
Game heo y amewo k o mi iga ing he cos pendulum
in public cons uc ion p ojec s
Games
P o ided in Coope a ion wi h:
MDPI – Mul idisciplina y Digi al Publishing Ins i u e, Basel
Sugges ed Ci a ion: Gia , Yahel; Mi elman, Amichai (2025) : Game heo y amewo k o mi iga ing
he cos pendulum in public cons uc ion p ojec s, Games, ISSN 2073-4336, MDPI, Basel, Vol. 16, Iss.
2, pp. 1-14,
h ps://doi.o g/10.3390/g16020011
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/330125
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
Academic Edi o : Kjell Hausken
Recei ed: 18 No embe 2024
Re ised: 21 Feb ua y 2025
Accep ed: 25 Feb ua y 2025
Published: 3 Ma ch 2025
Ci a ion: Gia , Y., & Mi elman, A.
(2025). Game Theo y F amewo k o
Mi iga ing he Cos Pendulum in
Public Cons uc ion P ojec s. Games,
16(2), 11. h ps://doi.o g/10.3390/
g16020011
Copy igh : © 2025 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license
(h ps://c ea i ecommons.o g/
licenses/by/4.0/).
A icle
Game Theo y F amewo k o Mi iga ing he Cos Pendulum in
Public Cons uc ion P ojec s
Yahel Gia 1,*,† and Amichai Mi elman 2,†
1Depa men o Indus ial Enginee ing, Je usalem College o Technology, Je usalem 9116001, Is ael
2Depa men o Ci il Enginee ing, A iel Uni e si y, A iel 4070000, Is ael; [email p o ec ed]
*Co espondence: [email p o ec ed]; Tel.: +972-26751165
†These au ho s con ibu ed equally o his wo k.
Abs ac : The coexis ence o he winne ’s cu se and cos o e uns in he cons uc ion
indus y implies a cos pendulum in which he winning bid is unde alued, whe eas he
inal paymen o he con ac o is o e alued. We posi ha his esul s om a s a egic
in e ac ion be ween h ee s akeholde s: he public agency (PA), he p ojec manage (PM),
and he winning con ac o , and we p opose a game- heo e ic amewo k o model his
dynamic. In he cu en s a e o p ac ice, he subgame be ween he con ac o and he
PM leads o oppo unis ic con ac o beha io and lenien supe ision, esul ing in in-
c eased cos s o he PA. We analyze how p ocedu al and cul u al in e en ions by he PA,
speci ically shi ing om a low-bid o an a e age-bid auc ion and incen i izing s ic e PM
o e sigh , al e he s a egic equilib ium. Ou indings indica e ha while each change
alone p o ides limi ed imp o emen , implemen ing bo h signi ican ly educes cos o e -
uns by aligning s akeholde incen i es. The indings o his analysis p o ide insigh in o
how public agencies can mi iga e he widesp ead p oblem o cos o e uns.
Keywo ds: p ojec managemen ; cons uc ion; cos o e uns; public ende s; oppo unism
1. In oduc ion
Cos o e uns in cons uc ion p ojec s ha e long been ecognized as a widesp ead
global phenomenon ha imposes signi ican cos s on socie y Thijssen (2022). Es ima es o
how many p ojec s expe ience cos o e uns ha e anged om 50% Lo e e al. (2019a) o
almos 90% Fly bje g e al. (2002). Despi e he many a emp s o ackle his eno mous bu -
den on budge s, i appea s ha p ac ical solu ions we e hus a ine ec i e in o e coming
i , as Lo e e al. (2022) lamen , “Hund eds o yea s o pain, wi h minimal gain”.
Wi h budge s cons ained by o e whelming p ojec cos s, policymake s a e s i ing o
educe cos s du ing ea lie phases, such as he ende ing phase. Indeed, low-bid ende s
a e he mos common o m o bidding o cons uc ion p ojec s in he public sec o , d i en
by he need o compe i ion, anspa ency, and equal oppo uni y Coggins e al. (2016);
El-Sayegh e al. (2021). A well-known ou come o low-bid ende s is he winne ’s cu se
phenomenon Capen e al. (1971), whe e he winning con ac o is cu sed by being obliga ed
o p o ide cons uc ion se ices a a sub-op imal p ice. Su p isingly, when examining em-
pi ical da a, i is ound ha con ac o s a e equen ly able o ecoup hei cos s du ing he
cons uc ion phase, esul ing in cos o e uns. Tha is, one may in e a pendulum e ec in
which a low winning bid pendula es o a inal in la ed paymen
Gia and Mi elman (2023)
.
In his pape , we model he s a egic in e ac ions be ween s akeholde s in he con-
s uc ion indus y. These s akeholde s a e he public agency (PA), p ojec manage s (PMs),
and cons uc ion con ac o s. By analyzing he payo s uc u e, ou model shows ha
Games 2025,16, 11 h ps://doi.o g/10.3390/g16020011
Games 2025,16, 11 2 o 14
he pendulum dynamic is he equilib ium ou come. In his p ocess, con ac o s engage in
oppo unis ic beha io and de ec i e s a egies, whe eas PMs adop a lenien managemen
s yle. The game be ween con ac o s and PMs, howe e , neglec s he PA, which is pe haps
he mos impo an s akeholde o he cons uc ion p ojec . The main hesis o his pape is
ha he PA can make cul u al and ope a ional changes ha will a ec he game be ween
PMs and con ac o s. These changes, i designed co ec ly, a e shown o be able o mi iga e
o e en emedia e he pendulum e ec .
The ex ended game is a wo-pe iod game in which he PA ac s in he i s pe iod,
whe eas he PM and he con ac o pa icipa e in subsequen pe iods. The e o e, he PA’s
op imal ac ion is such ha he ou come o he subsequen pe iods’ games esul s in he bes
alue o he PA. To ob ain his, he PA may ins i u e p ocedu al changes in he ende ing
p ocess ( ha mainly a ec he winning con ac o ’s ou come) and cul u al changes in how
he PM’s pe o mance is measu ed. We compa e how each o hese changes a ec s he PA’s
ou come and show ha while each change sepa a ely ha dly esul s in an imp o emen o
he PA, he ins i u ion o bo h changes does ha e a p o ound posi i e e ec . Thus, he game
heo y amewo k demons a es ha a obus s a egy o o e come he pendulum e ec
mus include mul iple measu es explaining he disappoin ing esul s o many a emp s
o o e come cos o e uns. The insigh s om his s udy can, he e o e, p o ide guidance
o indus y p ac i ione s and policymake s in de eloping mo e e ec i e s a egies o
cons uc ion p ojec managemen .
2. Li e a u e Re iew
2.1. Cons uc ion Tende s and Cos O e uns
Cos o e uns in cons uc ion p ojec s ha e been he subjec o many s udies, bo h
heo e ical and empi ical, wi h li le appa en gain Lo e e al. (2022). Resea che s ha e
employed di e en app oaches in an a emp o unde s and he causes o cos o e uns.
One line o in es iga ion is o obse e a pa icula indus y o manage ial mechanism
and glean he speci ic o echnical aspec s o ha indus y o mechanism ha esul in
cos o e uns. Fo example, in unneling p ojec s, Pa aske opoulou and Bou sis (2020)
showed ha inal cos s a e s ongly a ec ed by p elimina y g ound in es iga ions. In
ansi p ojec s, Gao and Tou an (2020) examined di e ences be ween hea y and ligh ail,
inding ha he o me su e ed conside ably mo e cos o e uns han he la e . Was y and
Nassa (2021) discussed p ojec delays and cos s emana ing om deba able issues be ween
he con ac o and he employe .
Ano he app oach is a sys ema ic obse a ion o cons uc ion p ojec s and en e p ises
h ough in e iews, wo kshops, and ques ionnai e su eys. The p oduc o his in es i-
ga ion was an in en o y o possible causes. Fo example, Famiyeh e al. (2017) su eyed
consul an s and con ac o s om six y go e nmen school p ojec s in Ghana. Simila ly,
Rosen eld (2014) analyzed esponses om wo hund ed p ac i ione s in he Is aeli cons uc-
ion indus y.
In pa allel o hese wo app oaches, ano he line o in es iga ion has eme ged, a guing
ha a mo e undamen al unde s anding o he o ces a play is needed. This app oach
is mo e s a egic in i s na u e, a o ing in ui i e easoning and ocusing on he s a egic
mo i a ion o s akeholde s o e he ac ical aspec s o he p ojec o he business en i on-
men . Two main heo ies ha e aken his cou se: he hiding hand hypo hesis da ing back o
Hi schman (1967) and he planning allacy, o iginally p oposed by T e sky and Kahneman
(1974) and la e championed by Fly bje g (2016) in he con ex o capi al p ojec s as a
coun e o he Hiding Hand.
Acco ding o he hiding hand hypo hesis, he planne s’ igno ance in he e y ea ly
s ages o he p ojec s imula es c ea i i y and p oblem-sol ing in he la e s ages o he
Games 2025,16, 11 3 o 14
p ojec . Thus, he hiding hand is a bene olen hand ha allows us o eap bene i s ha
would o he wise no be ealized. In con as , Fly bje g e al. (2002) a gued ha cos
o e uns in la ge capi al p ojec s a e he esul o in en ional mis ep esen a ion a he
easibili y s age, which lies wi h which p ojec p omo e s seek o con ince decision-make s
o ob ain p ojec app o al. Acco ding o his explana ion, he hiding hand is any hing bu
male olen and ins i u ional checks and balances a e needed o es ain i . In a simila
pessimis ic one, Loca elli e al. (2017) a gued ha he cons uc ion managemen li e a u e
igno es he elephan in he oom, i.e., co up ion. Fly bje g (2016)’s c i icism o he hiding
hand i sel was hea ily c i icized by Anheie (2016); Ika (2018); Lo e e al. (2019b). Al hough
Lo e e al. (2019a) ques ioned whe he ei he he hiding hand o he planning allacy a e
es able, hey a e epea edly used as ounda ional heo ies o imp o emen in p ojec
managemen Ika e al. (2020); Lo allo e al. (2023).
The s a egic beha io (o misbeha io ) in he planning s age is bu a p elude o
he s a egic in e ac ion in he ende ing s age. Low-bid ende s, pa icula ly in public
p ojec s, a e adi ionally he mos widesp ead bidding me hod o in as uc u e and
cons uc ion p ojec s, e.g., El-Sayegh e al. (2021); Onsa igo e al. (2020). The need o
gua an ee anspa ency, publici y, and equal oppo uni y in public p ocu emen demands
clea p ocedu es o be ollowed by bidde s, ende ing he low-bid me hod op imal om
a legal pe spec i e Awwad and Ammou y (2019). The low-bid me hod may esul in
highe cos s han he compe i i e p ice, such as when collusion happens Signo e al.
(2023). Howe e , his is ue wi h almos any bidding mechanism, and wi h p ope sc u iny,
collusion a emp s could be de ec ed Signo e al. (2020a).
No wi hs anding illegal collusion, low-bid ende s ha e been c i icized o ailing
o selec he bes alue bid Balles e os-Pé ez e al. (2015). In pa icula , low-bid ende s
equen ly esul in a phenomenon known as he winne ’s cu se. This was i s epo ed by
Capen e al. (1971), who ound ha auc ion winne s in he oil indus y consis en ly ecei ed
lowe e u ns on in es men compa ed o hei ini ial expec a ions. The winne ’s cu se has
since been ound in a e y wide ange o business se ings, including cons uc ion con ac s
(see Gia and Mi elman (2023)).
Consequen ly, al e na i e ende ing me hods ha e been p oposed. Fo example, addi-
ional c i e ia o he han he bes p ice ha e been adop ed by a ious public agencies whe e
a o mula inco po a ing c i e ia conce ning p ice, echnical, and o ganiza ional conside -
a ions de e mines he bes bid W. C. Wang e al. (2013). Ano he app oach o quali ying
he low-bid me hod is p esn ed by Signo e al. (2020b), who sugges ed ha only bids
abo e an “i esponsible bid h eshold” should be accep ed. Simila ly,
Che an e al. (2018)
p oposed a wo-s age auc ion p ocess wi h mul iple a ibu es ha allows he clien o
design an app op ia e bid-sco ing o mula. O he al e na i es include he second low-bid
me hod and he a e age-bid me hod. See, o example, Awwad and Ammou y (2019), who
ound ha om he clien ’s pe spec i e, he below-a e age-bid me hod is p e e ed o less
compe i i e p ojec s, whe eas o highly compe i i e p ojec s, he second low-bid me hod
is p e e ed. Balles e os-Pé ez e al. (2015) candidly concluded ha in he cu en s a e o
p ac ice, much is le o be desi ed as ende ing me hods a e based mo e on in ui ion han
on heo e ical and empi ical conside a ions.
2.2. P ojec Managemen in he Cons uc ion Indus y
I is un o una e ha despi e he g ea in es men in imp o ing p ojec managemen ,
p ojec manage s equen ly ail o deli e he p ojec as eques ed J. Va ajão and C uz-
Cunha (2013). Managing cons uc ion p ojec s, in pa icula , is a challenge due o he
e e -changing na u e o he p ojec s in ol ed Ziemba and Obł
a¸
k(2015). Acco ding o
Das and Rahman (2010), “oppo unism is de ined as a beha io by a pa y ha pu sues
Games 2025,16, 11 4 o 14
sel -in e es s wi h decei a he expense o o he pa ies”. Con ac o s’ oppo unis ic
beha io esul ing in losses o he clien could be conside able, and many esea che s
and p ac i ione s a e ad oca ing o a collabo a i e ela ionship be ween he clien and
he con ac o
Sup ap o e al. (2015)
. Key ac o s ha enable such coope a ion a e us ,
commi men , and eliabili y Deep e al. (2021). Thus, be e incen i es and means o
measu ing success ul pe o mance a e needed i clien s wish o change p ojec manage s’
a i udes o s ic ly en o ce he con ac J. E. Va ajão (2018).
Y. Wang e al. (2019) s udied how con ac s a e used o cu b oppo unism and dis in-
guished be ween “ha d” and “weak” o ms o oppo unism. While many s udies ocus
on he p e en ion o oppo unism, he e a e much ewe ha examine he en o cemen
o con ac s a e iola ions ook place. A ecen s udy o Chinese gene al con ac o s by
Yao e al. (2023)
examined h ee ypes o iola ions: le e iola ions, mu ually ag eed spi i
iola ions, and unila e ally assumed spi i iola ions.
2.3. Game Theo y Models
The bidding p ocess has been equen ly modeled using game heo y, e.g., by Budde
and Minne (2015); Lo en ziadis (2016). In pa icula , he winne ’s cu se phenomenon is
explained and analyzed using game heo y echniques (see Kagel and Le in (2002)). Fo
cons uc ion p ojec s, Ahmed e al. (2016) conside ed he compe i ion in mul is age bidding,
whe e he gene al con ac o who submi s he bids employs subcon ac o s o he ac ual
cons uc ion. Ba ough e al. (2012) used he p isone ’s dilemma and he chicken game o
acili a e he esolu ion o con lic s in cons uc ion p ojec s. The p isone ’s dilemma was
simila ly u ilized by Ho and Hsu (2014) o explain how o imp o e cons uc ion ende s ha
a e cha ac e ized by high bid p epa a ion cos s. Kembłowski e al. (2017) used no mal- o m
noncoope a i e games o model he bidding p ocess o a highway cons uc ion p ojec
om he bidde ’s pe spec i e.
Cos o e uns we e in es iga ed by Can a elli e al. (2013) using a game heo y
signaling model. They demons a ed ha misguided signaling leads o cos o e -
uns.
Peldschus e al. (2010)
used game heo y o selec an op imal cons uc ion si e.
San C is óbal (2015) used a model based on game heo y in o de o iden i y he ac i i ies
ha a e esponsible o he delays in cons uc ion p ojec s and o help guide PMs in dealing
wi h p ojec delays. All hese pape s used single-pe iod games. In con as , Shang and
Abdel Aziz (2020) de eloped a wo-pe iod game o assis public agencies in designing
paymen mechanisms o public–p i a e pa ne ship anspo a ion p ojec s.
The ubiqui y o cos o e uns in cons uc ion p ojec s sugges s ha hese ou comes
a e sys emic and oo ed in di e ing incen i es and payo s o he di e en s akeholde s
( he public agency, p ojec manage s, and he winning con ac o ). The e o e, game heo y
is an ideal ool o he analysis o his in e ac ion. No wi hs anding he game heo y s udies
desc ibed abo e, i appea s ha he po en ial o game heo y o explain cos o e uns in he
cons uc ion indus y is ye o be ealized. Indeed, a e iew by Pi a eenan (2019) o game
heo y applica ions concluded ha heo e ic game analyses o cons uc ion p ojec manage-
men ha e g ea po en ial bu a e cu en ly nascen . E en mo e ecen ly,
Eissa e al. (2021)
conduc ed a e iew o game heo y applica ions in he ield o cons uc ion enginee ing
and concluded ha “e en hough he cons uc ion indus y is ich wi h encoun e s among
i s s akeholde s, cons uc ion enginee ing and managemen esea ch lacks a ho ough
in es iga ion o game- heo e ic applica ions”. Simila ly, Na bae e al. (2022) conduc ed
a ho ough e iew o game heo y models in p ojec managemen and ound ha he
li e a u e “is no as ich as he ones in he o he a eas o business managemen , such as
inance, ope a ions managemen , and supply chain managemen ”. Fu he mo e, no one
Games 2025,16, 11 5 o 14
o he s udies hey e iewed conside ed he mul i-le el s a egic ela ionship be ween he
public agency, p ojec manage s, and he con ac o . This gap s udy a emp s o ill ha gap.
3. Model Backg ound
Ou model builds on a case s udy by Gia and Mi elman (2023) conduc ed in he
Depa men o Enginee ing and Cons uc ion (DEC), a la ge public o ganiza ion in Is ael.
The DEC employs PMs who o e see he cons uc ion o p ojec s by con ac o s a e a
low-bid ende ing p ocess. They analyzed a da a se o 148 low-bid cons uc ion p ojec s
and con i med he pendulum e ec desc ibed abo e: on a e age, he winning bid was 8%
lowe han he expe cos es ima e, whe eas he inal paymen o he con ac o was 16%
g ea e han he cos es ima e.
Depending on p ojec complexi y, manage ial s uc u es in DEC can di e signi ican ly,
encompassing a ange o di e se membe s, posi ions, esponsibili ies, and hie a chies.
Ne e heless, he pendulum e ec emained consis en ac oss di e en p ojec complexi ies.
The e o e, i is a gued ha he s a egic in e ac ion can be simpli ied o h ee membe s:
PMs, con ac o s, and he PA. Ul ima ely, he PM’s app o al is necessa y o he con ac o
o be ewa ded beyond hei binding o e . A guably, such decisions a e shaped by he
egula ions and no ms es ablished by he PA, he p essu es imposed by he con ac o , and
he incen i es and pe sonali y o he PM.
Löws ed and Räisänen (2012) emphasized he impo ance o quali a i e s udies and
subjec i e na a i es in unde s anding o ganiza ional beha io . The e o e, we complemen
he backg ound o ou model wi h in o mal subjec i e na a i es om expe ienced senio -
le el PMs om DEC. Via a social app, we sen he ollowing ques ion:
In he cons uc ion indus y, he ollowing global phenomenon has been iden i ied: con ac o s
win ende s wi h o e s ha a e below ma ke alue bu , a he comple ion o he p ojec , a e ewa ded
abo e ma ke alue. How do you explain his phenomenon?
Below a e exce p s om answe s ha we ecei ed:
•
“In he beginning o a p ojec , all in ol ed in he p ojec a e awa e o he eali y ha
he winning bid is well below ma ke alue. This unde s anding makes i easie o
PMs and hei supe iso s o app o e addi ional claims. This dynamic esul s in he
con ac o being o e paid”.
• “Con ac o s a e ained o ind loopholes in ende s, and will bid low assuming ha
hey will be able o exploi hese o hei a o du ing p ojec execu ion. The p oblem
is, e en when he design is op imal and igo ous, changes in p ojec s a e una oidable.
Ul ima ely, a leas some o he claims a e jus , o in he g ey a ea. Mos PMs will yield
o he con ac o s’ p essu es and app o e hei claims”.
•
“The cons uc ion indus y has in o mally accep ed cos o e uns as a eali y, and
highe managemen is ocused on ge ing p ojec s comple ed. In some cases, usually
wi h young PMs wi h li le expe ience, hey will play ough wi h he con ac o .
Howe e , a some poin , hey will con o m o he no m and app o e special claims
made by he con ac o . Ei he because hey unde s and ha o he wise hey will ail
o deli e he p ojec , o because highe managemen will o e ule hei decisions”.
•
“The p i a e sec o selec s con ac o s acco ding o hei p e ious epu a ion. Public
ende s use he low bid me hod and he e o e con ac o s a e obliged o submi
un ealis ic bids. I he PM will be s ic in e ms o budge , hen ei he he p ojec
quali y would be comp omised, o he schedule would be delayed. The con ac o
will no pay ou -o -pocke unde any scena io”.
•
“This eali y exis s no only in he cons uc ion indus y. E e y h ee yea s, ca e ing
con ac o s in ou o ganiza ions a e also hi ed ia low bid ende s. We no iced a
Games 2025,16, 11 6 o 14
pa e n we e he winning con ac o g adually comp omised he quali y o he meals
se ed. I imagine his is he same ype o ick ha building con ac o s use”.
PMs we e hen asked o espond o a second and inal ques ion:
Wha is he in luence o a lenien e sus s ic app oach when managing a con ac o in
a p ojec ?
The ollowing a e exce p s om esponses o his ques ion:
•
“A s ic app oach can be bene icial in si ua ions whe e he con ac o has a his o y
o unde pe o ming o aking ad an age o lenien PMs. I can se e as a de e en
agains cos o e uns and ensu e ha he p ojec adhe es o he ini ial plan. Howe e ,
in some cases he PMs app oach has a limi ed impac on p ojec ou comes”.
•
“A lenien app oach may os e be e collabo a ion be ween he PM and he con ac o ,
leading o a mo e posi i e wo king en i onmen . Howe e , i could also esul in cos
o e uns and schedule delays, as he con ac o migh ake ad an age o he PM’s
lexibili y o eques addi ional unds o ime”.
•
“A s ic app oach can help main ain con ol o e he p ojec budge and imeline,
minimizing he isk o cos o e uns and delays. Howe e , an o e ly igid app oach
may s ain he ela ionship be ween he PM and he con ac o , po en ially leading o
dispu es, claims, and educed o e all p ojec quali y”.
•
“A balance be ween leniency and s ic ness is c ucial. A PM should be lexible enough
o accommoda e necessa y changes o unexpec ed challenges, while s ill ensu ing ha
he con ac o is held accoun able o deli e ing he p ojec wi hin he ag eed-upon
budge and imeline. This app oach encou ages collabo a ion and p oblem-sol ing
while main aining he in eg i y o he p ojec ”.
•
“A lenien app oach migh esul in a mo e adap i e and esponsi e p ojec eam ha
can quickly add ess unexpec ed issues o changes. Howe e , i could also make i
challenging o he PM o en o ce con ac ual obliga ions, po en ially jeopa dizing he
p ojec ’s success”.
To summa ize, he senio manage s’ esponses sugges a s a egic line o easoning
o he exis ence o he pendulum e ec . Thei esponses show ha hey belie e he PMs’
choice o app oach, s ic o lenien , has a subs an ial impac on PM–con ac o in e ac ion.
O e all, his quali a i e s udy con ibu ed o ou unde s anding o he p ojec dynamics
om he PMs’ pe spec i e and acili a ed he de elopmen o he game heo y model.
4. Game Theo y F amewo k
4.1. Basic Model: The Cu en S a e
The e en s ollowing he con ac o ’s winning o he low bid ende can be modeled
as a simul aneous noncoope a i e game be ween wo playe s: he con ac o and he PM.
In he simples o m, each playe has exac ly wo possible ac ions wi h espec o hei
commi men o he con ac be ween he PA and he con ac o . The pendulum e ec ha
was desc ibed in he p e ious sec ions is a gued o be he esul o a s a egic in e ac ion
be ween he wo playe s.
Fo he con ac o ’s ac ions, we used he e minology o he classical p isone ’s
dilemma game: commi (C) and de ec (D). The con ac o may choose o “commi ”,
i.e., o hono all he con ac ’s s ipula ions and sa is ac o ily add ess all he issues aised by
he PM. Al e na i ely, he con ac o may choose o “de ec ”, meaning ha he con ac o
will use a ious ac ics ha will minimize hei cos s and maximize hei e enue Das and
Rahman (2010); Liu e al. (2016); Y. Wang e al. (2019). Examples o such oppo unis ic
beha io include he ollowing:
• Comp omising quali y, sa e y, and/o schedule.
Games 2025,16, 11 7 o 14
• Sea ching o loopholes in he binding con ac .
• Engaging in dispu es (e.g., making alse claims, iling lawsui s, e c.).
• Using co up means o a oid ul illing obliga ions.
The PM is cha ged by hei employe (i.e., he PA) o ep esen he depa men and
ensu e ha he p ojec is execu ed in s ic compliance wi h he con ac ’s e ms. The e o e,
i is he du y o he PM o con on he con ac o whene e needed. Howe e , being
con on a ional does no come wi hou cos s o he PM. Fi s , a s ic app oach akes an
emo ional oll on he PM. Second, s ic ness is usually me wi h con on a ion. Since he
PA’s cul u al wo kplace clima e is con on a ion-a e se, manage s do no ewa d PMs
o s ic ness. Indeed, i has been acknowledged in he li e a u e ha he e a e e o s
in he cons uc ion indus y o shi om ad e sa ial ela ionships o mo e collabo a i e
ones Pa ucco e al. (2021). Finally, since cos o e uns a e so common, he PM bea s no
epu a ion cos s when he p ojec unde hei managemen exceeds i s budge . Acco dingly,
he PM’s possible ac ions a e o be “s ic ” (S) o o be “lenien ” (L). Leniency implies
ha he PM does no con on he con ac o when he la e beha es unsc upulously.
Fu he mo e, leniency also implies negligence ha ad e sely a ec s compliance wi h
he con ac e ms e en when he con ac o does gene ally ollow he con ac (i.e., he
con ac o ac s owa ds C).
Figu e 1desc ibes he simul aneous game be ween he PM and he con ac o in
no mal o m. We assumed ha he playe s a e isk-neu al and, he e o e, payo s can be
iewed as he means o he ou comes. The payo s o he con ac o a e p esen ed as a
pe cen age o he ue cos . By ue cos , we mean he ope a ional cos s (e.g. ma e ials, labo ,
e c.) plus he indus y’s ypical ma gin o p o i .
Ve sion Feb ua y 21, 2025 submi ed o Games 6 o 13
PM
S ic (S) Lenien (L)
Con ac o Commi (C) (−δb,h) (−δb+ϵ,h−)
De ec (D) (δp−ϵ−x,l) (δp,m)
Figu e 1. The game be ween he con ac o and he PM in he cu en si ua ion.
4. Game Theo y F amewo k 248
4.1. Basic Model: The Cu en S a e 249
The e en s ollowing he con ac o ’s winning o he low bid ende can be modelled
250
as a simul aneous noncoope a i e game be ween wo playe s: he con ac o and he PM.
251
In he simples o m, each playe has exac ly wo possible ac ions wi h espec o hei
252
commi men o he con ac be ween he PA and he con ac o . The pendulum e ec ha
253
was desc ibed in he p e ious sec ions is a gued o be he esul o a s a egic in e ac ion
254
be ween he wo playe s. 255
Fo he con ac o ’s ac ions we used he e minology o he classical P isone ’s
256
dilemma game, commi (C) and de ec (D). The con ac o may choose o “commi ", i.e., o
257
hono all he con ac ’s s ipula ions and sa is ac o ily add ess all he issues aised by he
258
PM. Al e na i ely, he con ac o may choose o “de ec ", meaning ha he con ac o will
259
use a ious ac ics ha will minimize hei cos s and maximize hei e enue [
33
,
37
,
54
].
260
Examples o such oppo unis ic beha io include: 261
• Comp omising quali y, sa e y, and/o schedule. 262
• Sea ching o loopholes in he binding con ac . 263
• Engaging in dispu es (e.g., making alse claims, iling lawsui s, e c.). 264
• Using co up means o a oid ul illing obliga ions. 265
The PM is cha ged by hei employe (i.e., he PA) o ep esen he depa men and
266
ensu e ha he p ojec is execu ed in s ic compliance wi h he con ac ’s e ms. I is
267
he e o e he du y o he PM o con on he con ac o whene e needed. Howe e , being
268
con on a ional does no come wi hou cos s o he PM. Fi s , a s ic app oach akes an
269
emo ional oll on he PM. Second, s ic ness is usually me wi h con on a ion. Since he
270
PA’s cul u al wo kplace clima e is con on a ion-a e se, manage s do no ewa d PMs o 271
s ic ness. Indeed, i has been acknowledged in he li e a u e ha he e a e e o s in he
272
cons uc ion indus y o shi om ad e sa ial ela ionships o mo e collabo a i e ones
273
[
55
]. Finally, since cos o e uns a e so common, he PM bea s no epu a ion cos s when
274
he p ojec unde hei managemen exceeds i s budge . Acco dingly, he PM’s possible
275
ac ions a e o be “s ic " (S), o o be “lenien " (L). Leniency implies ha he PM does no
276
con on he con ac o when he la e beha es unsc upulously. Fu he mo e, leniency also
277
implies negligence ha ad e sely a ec s compliance wi h he con ac e ms e en when
278
he con ac o does gene ally ollow he con ac (i.e., he con ac o ac s C). 279
Figu e 1desc ibes he simul aneous game be ween he PM and he con ac o in
280
no mal o m. We assumed ha he playe s a e isk-neu al and he e o e he payo s can
281
be iewed as he means o he ou comes. The payo s o he con ac o a e gi en as a
282
pe cen age o he ue cos . By ue cos , we mean he ope a ional cos s (e.g. ma e ials, labo ,
283
e c.) plus he indus y’s ypical ma gin o p o i . 284
We le
δb
deno e he ela i e di e ence be ween he winning bid and he ue cos ,
285
and le
δp
deno e he ela i e di e ence be ween he paymen o he con ac o and he
286
es ima ed cos . The a iable
δb
e lec s he unde paymen o he con ac o and he sa ings
287
o he PA, whe eas
δp
e lec s he cos o e uns a e. Fo example, in he case s udy [
8
]
288
desc ibed in he p e ious sec ion, he winning bid was 92% o he es ima ed cos , and
289
acco dingly
δb=
8%. The paymen s o he con ac o we e 116% o he es ima ed cos and
290
he e o e in ha case δp=16%. 291
Figu e 1. The game be ween he con ac o and he PM in he cu en si ua ion.
We le
δb
deno e he ela i e di e ence be ween he winning bid and he ue cos
and le
δp
deno e he ela i e di e ence be ween he paymen o he con ac o and he
es ima ed cos . The a iable
δb
e lec s he unde paymen o he con ac o and he sa ings
o he PA, whe eas
δp
e lec s he cos o e un a e. Fo example, in he case s udy by Gia
and Mi elman (2023) desc ibed in he p e ious sec ion, he winning bid was 92% o he
es ima ed cos , and acco dingly,
δb=
8%. The paymen s o he con ac o we e 116% o he
es ima ed cos and, he e o e, in ha case, δp=16%.
No ice ha when we e e o he con ac being unde paid o o e paid, i is ela i e
o he ue cos and no he winning bid. This aci ly assumes ha he con ac o ’s p o i
expec a ions a e simila o he ypical indus y p o i ma gin. Howe e , his assump ion
is made mainly o con enience, and he ensuing analysis will no ma e ially change i
he con ac o ’s p o i expec a ions do no di e signi ican ly om he indus y-accep ed
p o i ma gins.
Payo s o he PM a e o dinal and a e deno ed by
h
,
h−
,
m
,
l
, whe e
h>h−>> m>l
.
We now explain he payo s o he con ac o and he PM o each ou come o he game,
whe e ou come (X,Y) deno es he con ac o playing X and he PM playing Y.
•
Ou come (C,S)— he con ac o commi s and he PM is s ic : Since he con ac e ms
a e ollowed, he con ac o ends up wi h a
δb
unde paymen . Since he commi men
was ob ained wi hou con on a ion, he PM’s payo is high (h).
•
Ou come (C,L)— he con ac o commi s and he PM is lenien : Al hough he con ac o
is commi ed o he con ac , he PM’s leniency esul s in g ea e paymen s o he
Games 2025,16, 11 8 o 14
con ac o han when (C,S) is played. We assume ha he di e ence be ween a lenien
and s ic PM amoun s o a di e ence o
ϵ>
0 pe cen age poin s. Tha is,
ϵ
ep esen s
how much he paymen o he con ac o inc eases i a s ic PM changes hei beha io
o lenien . The PM’s payo is sligh ly lowe han in he p e ious case (
h−
) since he
ex a
ϵ
o he con ac o is a he expense o he PA and is a ibu ed o he PM’s
slack beha io .
•
Ou come (D,L)— he con ac o de ec s and he PM is lenien : The con ac o ’s un-
sc upulous an ics a e le unchecked by he PM; he e o e, he con ac o is o e paid
δp
. Due o he o ganiza ion’s cul u e, he PM’s payo is medium (
m
) since, al hough
he e a e conside able p ojec o e uns, hese a e somewha ole a ed when he e a e
no PM–con ac o con lic s.
•
Ou come (D,S)— he con ac o de ec s and he PM is s ic : The PM’s s ic ness
dec eases he paymen s o he con ac o by
ϵ
. In addi ion, he con ac o has expenses
ha a e beyond no mal cons uc ion- ela ed cos s (e.g., addi ional manage ial and
legal cos s). These a e assumed o be
x
pe cen age poin s. The e o e, he con ac o ’s
ne o e paymen is
δp−ϵ−x
. These ex e nal cos s a e ne e oo la ge o de e
de ec ion, namely ha
δp−ϵ−x>−δb
and, he e o e,
x<δb+δp−ϵ
. The PM’s
payo is low (
l
) because o he o ganiza ion’s cul u e ha does no ewa d s ic ness i
i leads o con on a ions.
No e ha
ϵ
ep esen s he added paymen s o he con ac o as a esul o he PM’s
leniency. We easonably assumed ha i could no be mo e han he maximal o e uns, i.e.,
ϵ<δp.
Ou come (D,L) is he unique Nash equilib ium o he game. Tha is, he con ac o is
decei ul, and he PM yields o he con ac o . The e o e, in equilib ium, he con ac o is
o e paid
δp
. The con ac o ’s o e paymen is a mani es a ion o he pendulum e ec . I
does no only elimina e he an icipa ed unde paymen o he low bid bu e en e e ses he
inal ou come o esul in an o e paid con ac o .
4.2. Public U ili y
F om he PA’s (i.e., he public) pe spec i e, he equilib ium ou come is he wo s
possible ou come. Empi ical e idence sugges s ha his is equen ly he de ac o esul
in public agencies Gia and Mi elman (2023). The unde lying o ce d i ing his esul is
ha he con ac o is always mo i a ed o de ec . Consequen ly, i is be e o he PM o be
lenien and a oid con on a ion.
The cu en si ua ion appea s o bene i he con ac o a he expense o he PA. This is
because he PA’s ne gain om he p ojec is gene ally he addi i e in e se o he con ac o ’s
o e paymen . Fo example, i he con ac o is o e paid
δp
(such as o (D,L)), hen he PA’s
losses a e
δp
. The only excep ion is he (D,S) ou come. He e, he PA’s loss is only
δp−ϵ
whe eas he con ac o ’s ne o e paymen is
δp−ϵ−x
because
x
ep esen s ex e nal cos s
o he con ac o ha a e no a he expense o he PA.
5. The Ex ended Model: Mi iga ing Cos O e uns
The analysis o he cu en s a e e eals ha while he PM is he PA’s agen , hei
in e es s a e no aligned. The e o e, we need o b oaden he game o include he PA
as a hi d playe . We p opose ha he PA conside s cul u al and p ocedu al changes
o i s ope a ions ha will s a egically a ec he o he playe s’ ac ions and imp o e he
public’s u ili y.
•
Cul u al change: The cu en wo kplace cul u e is con on a ion-ad e se; he e o e,
i is un ewa ding o s ic PMs. An al e na i e app oach is o implemen ways o
measu e PMs’ pe o mance and ewa d hem based on he ac ual p ojec ou come
