Post on 07-Jul-2018
8/18/2019 Bbswp10002 Ok
1/43
8/18/2019 Bbswp10002 Ok
2/43
2
The persis tence of bank profit1
Abstract
This paper exam ines the s t r ength of compet i t ion in 65 n a t iona l bankin g indu st r ies .
Coun tr y-level dyn am ic pan el est imates of the persisten ce of bank pr ofi t are r eported and
comp ar ed. The persisten ce of bank pr ofi t appear s to be weaker for bank s in developing
coun tr ies th an for those in deve loped count r ies . Pers is tence i s re la t ively h ig h in Nort h
America and Western Eu r ope an d relat ively low in East Asia, the Pacific and Su b-Sahar an
Afr ica . The per s is tence of profi t is s t r onger when ent r y barr iers a re h igh , an d wh en
comp eti t ion is low accor ding to both st r uctu r e- an d cond u ct-based competi t ion in dicators.
Keywords: Bank ing , competi t ion, dyn am ic pan el est imation, entr y, profi tabil ity,
per s is ten ce
J EL: G21 L11
1 Special thanks to Barbara Casu, Francesco Cesarini, Franco Fiordelisi, Claudia Girardone, Roberto Violi and Giuseppe Zadra for detailedcomments on the paper. Also thanks to participants at the ABI Einaudi Workshop on European Bank Competition held at ABI Piazza delgesu 49, 18 September 20009, Rome. All errors and omissions as usual rest with the authors. The financial support of the Istituto (formerly,Ente) Luigi Einaudi, sponsoring the research project “Competition in European Banking” is gratefully acknowledged”
8/18/2019 Bbswp10002 Ok
3/43
3
The persis tence of bank profit
1. Introduction
The compet it ive land scape of banking has been t r ans form ed over the past th ree
decades by for ces su ch as der egu lation, techn ological cha ng e and th e globalization of goods
an d finan cial mar k ets. These developm ents h ave imp acted upon t he operations , efficiency,
pr od u ct iv it y , m ar g in s an d p r ofit abi li ty of ba n k s in a ll cou n tr ies (Dem ir g u c-Ku n t an d
Hu izan ga , 1998; Claessens et a l . , 1998; Levin e, 2003; Clark e et a l . , 2003; Car bo and
Fernan dez, 2007; Maudos an d Fern anad ez de Guevara , 2004; 2007; Berg er , 2007)2. Bank s
nowadays opera te in pr oduct and g eogr aphica l mar ke ts very d i f ferent f rom th ose tha t
exis ted th i r ty year s ag o. Bank s u se a r ang e of convent iona l an d inn ovat ive technologies to
serve the ir customers .
To date, most academic r esearch on competi t ion and i ts effects on ban k per for ma nce
is based on th eor etical models that a r e stat ic in n atu r e (Claessens an d Laeven, 2004; Sh affer
2004; Goddar d an d Wilson, 2009; Dick an d Han na n, 2009). Static models are us eful in
ident i fy ing causa l r e la t ionsh ips between key var iables when mar k e ts a re in equi libr ium ,
bu t th ey pr ovid e on ly a s n apsh ot p ic tu r e of a dy n am ic com pet it ive p r oces s (Ger os k i, 1990).
There is n o cer ta in ty th a t cond uct or perform ance measur es observed a t any poin t in t im e
represent equil ibrium values. For example, an empirical association between high
concent ra tion an d h igh pr ofi tabil ity (implici t in SCP-based models) may s imply appear by
chance , f rom observa t ions ta ken d ur ing a per iod wh en th e re levant mar ke t i s in a s ta te of
disequil ibriu m. If so, cross-sectional data does not captur e (un less by luck) th e long -ru n
equi libr iu m re la t ionship . Fu r th ermor e , cross-sect iona l da t a u sua l ly does n ot conta in
su fficient inform ation on wh ich to base r eliable policy decisions to pr omote competi t ive
2 Berger et al . (1995) and Jones and Cri tchfield (2005) provide an exhau st ive analysis of s tructur al and s trateg ic
chang e in US banking . Goddard et al . (2007) provide an overview of European ban king . DeYoung et al . (2004) and
Berg er et al . (2004) analyze the response of banks to chang es in th eir competit ive environm ent. Vander Vennet
(2002), St iroh and Str ahan (2003) and Maudos and Fern andez de Guevara (20 04) assess the impact of these
chang es on ban k performan ce. Cetorel l i (2004) exam ines the effect of bank competi t ion on th e s ize dis tr ibut ion of
non-finan cial fi rms. Beck et a l . (2004) and Cetorel l i and Str ahan (2006) analyse the effect of bank competi t ion on
the avai labi li ty of credi t to sm all, medium and large fi rm s. Black an d Strah an (2002), Bonaccorsi Di Pat ta an d
Dell‟Aricca (2004) an d Zarutsk ie (2003) assess the implicat ions for borr owing , investmen t , entr epreneur ship an dnew firm creat ion.
8/18/2019 Bbswp10002 Ok
4/43
8/18/2019 Bbswp10002 Ok
5/43
5
methodology th a t i s preva lent in m uch of the l ite ra tur e based on the s t r uc tur e condu ct
per for m an ce pa r ad ig m (S CP) an d th e New In du s t r ia l Or g an iza t ion (N EIO) li ter a tu r e.3
In i ts str ong est form , the persisten ce of pr ofi t hyp othesis developed by Mueller
(1977, 1986) con sists of two cond it ions . Fir st , entr y an d exit ar e sufficiently free to
elimina te any fir m‟s abn orm al pr ofit qu ickly. Second, al l firm s‟ pr ofi t r ates tend to
converg e toward s an ident ica l long -ru n averag e va lue . The second condi t ion is s t r onger
th an t h e first ; and u nd er a less r estrict ive ver sion of th e per sistence of profi ts hypoth esis,
abnormal pr ofi t ra tes d iss ipa te qu ickly an d converg ence of prof it r a tes i s toward s long-ru n
averag e values tha t may d iffer between fir m s. Th e emp irical an alysis in th is paper
concent ra tes upon th is la t ter form ulation of th e persisten ce of pr ofi t hypoth esis. The
a l tern a t ive hypothesis i s tha t s ome incum bent f irm s enjoy r egula tory pr otec t ion , or p ossess
th e capabil ity to pr event imitat ion or block entr y. If so, abnor m al profi t tends to persist
f rom year to year , and converg ence is e ither s low or , in th e most ext r eme case , non -
existent .
A nu mber of s tudies ha ve tes ted the pers is tence of pr ofi t s hyp othesis for th e (non-
ba n k ) m an u fa ct u r in g an d ser vices sector s (Mu el ler ; 1 986; Ger os k i an d J acq u em in , 1988;
Warin g , 1996 Goddar d an d Wilson, 1996, 1999; McGahan an d Port er, 1999; Glen et a l .,
2001, 2003; Maray um a an d Odagir i 2002) . This l i tera t ur e typica lly observes pa t te rn s in th e
time-series variat ion of firm -level pr ofit r ate data , an d dr aws infer ences about th e natu re of
compet i t ion (whether barr iers to ent r y exis t , and i f so, whether th ey a re temporary or
per m an en t). Mos t s tu d ies r ep or t an ave r ag e per s is t en ce of p r of it es t im ate, in th e for m of a
fir st-ord er au toreg ressive coefficient in a t im e-series r egr ession of firm -level norm alized
pr of it r a t es , in th e r an g e 0 .4 to 0 .5 . Som e stu d ies r ep or t ei th er lower or h ig h er va lu es .4
The
per s is ten ce of p r of it h as been fou n d to be lower in develop in g cou n tr ie s th an in develop ed
coun tr ies . This pa t te rn i s a t t r ibuted to lower sun k costs of ent ry , fas te r economic growth ,
th e role of govern m ent, and th e pr esence of larg e cong lomerat e firm s in developing
3 For an overview, see Berg er (1995), Berger et al . (2004), Dick and Hann an (2009), Goddard et al . (2001, 2004b),
Claessens and Laeven (2004), Shaffer (2004).
4 Goddard a nd Wilson (1999) an d Lipczyns ki et al. (2009) review evidence on the per sistence of profit in
manufacturing and services .
8/18/2019 Bbswp10002 Ok
6/43
6
coun tr ies wh ich ten d to incr ease th e in tensity of competi t ion experienced by domestic firm s
(Glen et a l . , 2001; Steph an an d Tsapin , 2008).
The evidence for th e persistence of pr ofi t in bank ing is r elat ively scant.5 Using a
sam ple of larg e US ban k s observed over t he period 1986 -91, Levonian (1993) find s th at
despite res t r a in ts on compet i t ion imposed by bank regu la t ion , abnormal pr ofi t s tend to be
temporar y , ra th er th an perman ent . The pace of converg ence, however , i s s lower th an t ha t
sug gested by most man ufac tur ing s tudies . Roland (1997) exam ines the pers is tence of
pr of it s of US ba n k h old in g com pan ies , u s in g qu ar t er ly da ta fo r th e per io d 1986 -92. Th e
per s is ten ce of p r of it is s t r on g er fo r ba n k s wit h be low -aver ag e per fo r m an ce . Ber g er et a l.
(2000) use a non-par am etric meth odology to exam ine the persisten ce of bank pr ofi t . The
str eng th of pers istence is foun d to differ between bank s init ia l ly located in th e top an d
bot tom decil es of th e d is t r ibu t ion of ba n k s by per for m an ce.
Goddar d et a l . (2004a) est imate th e pers isten ce of pr ofi t for a s am ple of Eur opean
ba n k s fr om six cou n t r ie s , u sin g a m od el th at in cor por a tes ba n k -s pecif ic va r ia bles in clu din g
size, diver sification, r isk an d owner sh ip type. The persisten ce of pr ofi t is h igh er for mu tu al
(savings and coopera t ive) bank s th an for comm erc ia l banks. By coun try , pers is tence i s
h igh est for Fran ce, where a s t rong regu la tory t r adi t ion may ha ve insula ted bank s f rom the
full r igou rs of comp eti t ion.
Sever al recent stu dies focu s on ind ividual count r ies. Agostt ino et a l. (2005 ) report
est imates of the persisten ce of pr ofit for I ta l ian bank s for th e period 1997 -2000. Pers istence
is posit ively associated with own ersh ip concentr ation. Knap p et a l . (2006) report
per s is ten ce es t im ates fo r a sam ple of US ba n k s , s u g g es t in g th a t p r of it s ta k e a s lon g a s five
years to converg e towar ds averag e indu st ry nor ms. Pers is tence est imates for Turk ey and
Greece are r eported by Bek tas (200 7) and Ath an asog lou et a l . (2008), respectively. In a
recent c ross-coun try s tu dy for Su b-Sahar an Afr ica, Flamini (2009) f inds tha t s t ron g
per s is ten ce is a ssocia ted pos it ively wit h ba n k s ize, d iver s if ic a t ion an d pr iva te ow n er sh ip .
Overall , the emp irical eviden ce to date focuses on a r elat ively sm all nu m ber of
coun tr ies, and iden tifies posit ive au tocorr elat ion in ban k p r ofit r ates obser ved over t im e.
5 St i roh and Strah an (2003) tes t the effect iveness of deregu lat ion in improving the competi t ive process .
Deregulat ion increases both en try and ex i t , and r eal locates mar ket sh are fr om un profi table bank s to profi table ba n k s .
8/18/2019 Bbswp10002 Ok
7/43
7
The persisten ce of bank pr ofi t is dr iven by bank -specific an d indu str y char acterist ics, an d
ma croeconomic cond it ions . Sur pr ising ly, there is l it t le evidence on th e ext ent to wh ich
ba n k r eg u la t ion an d su per vis ion d et er m in e d if fe r en ces in th e per s is t en ce of p r of it be tween
coun tr ies. A prin cipal objective of this pa per is to fi ll this g ap in t h e li teratu re.
3. Methods
This section describes th e emp irical m ethods u sed in this pap er. A two-stag e
approach is u sed to measur e the pers is tence of profi t , and invest iga t e the factors th a t
determ ine the persistence of pr ofit . At th e fir st stag e, dyn am ic pan el f ir st -order
au toreg r essive models ar e est imat ed for th e norm alized profit ra tes of ban k s in each of th e
65 coun tr ies. At the second stag e, th e first -order au toreg r essive coefficien ts obtained at th e
f irs t s tag e a re u sed as th e dependent var iable in a c ross-sec t iona l regr ession for th e
de term inants of the pers is tence of profi t in each countr y .
Stage 1 Est im at ing the persis tence o f prof it
In ord er to el imin ate th e effect of cyclical f lu ctuat ion s wh ich impact sim ilarly on th e
pr of it r a t es of a ll of th e ba n k s wit h in ea ch cou n tr y , i,t
is exp r essed as a deviation from th e
cross-sectional mean p rofi t rate in year t . Therefore i,t
i s th e norm al ized pr ofi t ra t e of bank
i in y ear t . In th e theoretical model, th e ch an g e in t he n orm alized profi t rate of bank i
between yea r t – 1 an d year t , denoted i,t
, is ass u med to be a fu nction of bank i‟s lag ged
norm al ized pr ofi t ra te ; the compon ent of cur rent and p ast en t r y wh ich im pacts d irec t ly on
ba n k i‟s pr ofit abi li ty , den ot ed Ei,t – k
for k =0 ... ; an d an idiosyncr atic compon ent, denoted ui,t
:
i,t
= i +
0k k t,ik , j E + u i,t (1)
8/18/2019 Bbswp10002 Ok
8/43
8
In (1), th e coefficient s j,k
, which measur e the imp act of Ei,t – k
on i,t
, a re assu med to be the
same for a l l bank s th a t a r e loca ted in coun try j . Ent ry i s assu med to be a fun c t ion of past
realizations of bank i‟s n orm alized profi t ra te:
Ei,t
= +
1k k t,ik , j + e i,t (2)
Subst itu t in g (2) in to (1) and re - par am et er iz in g y ie ld s an au tor eg r es sive m od el fo r ba n k i‟s
normalized profi t rate:
i,t
= t,i1k
k t,ik , ji v~
(3)
When employing pan e l da ta wi th a sh or t t im e-dimension , it i s convenient to es t imate a
fir st-ord er au toreg ressive (AR(1)) specification for i,t
, with t he h igher -order lagg ed profi t
ra tes supp ressed:
t,i1t,i jit,i v~ (4)
In (4), i~ denotes ban k i‟s long -ru n m ean norm al ized prof it ra te , and
j in (4) r eplaces
j,1 in
(3). The adjust men t of n orm alized ban k pr ofit ra tes described by (4) is inter pr eted as a
cons equence of the int eraction between pr ofi tabil i ty and th e en tr y th r eat , as postu lated in
th e contestable ma r kets l i teratu r e (Baum ol et a l. , 1982; Bratlan d, 2004).
Estim ation of th e pers istence of profi t coefficient s, j in (4), is implemen ted usin g Arellan o
an d Bover‟s (1995) sy stem GMM estimator , including both lagg ed differences an d levels of
the explana tory var iables as in st r um ents . The system GMM est imator redu ces potent ia l
biases in fin it e sam ples , an d asym ptot ic im p r ecis ion a ssoc ia t ed wit h Ar ellan o an d Bon d‟s
(1991) differen ced GMM estima tor (Blun dell an d Bond, 1998). The consistency of the
system GMM est imator depends both on the va lid ity of the assu mpt ion th a t th e er ror te rm
8/18/2019 Bbswp10002 Ok
9/43
9
is not au tocor related, and on th e validity of the instr u men ts. Two specification t ests ar e
repor ted . The f i rs t i s th e Hansen (1982) tes t , which is commonly u sed as a s tan dard tes t of
inst r um ent va lid ity .6 I t can also be viewed as a test of str u ctur al sp ecification (Roodm an ,
2006, 2008). Omitt ing impor tan t explan atory var iables, for exa mp le, could move
components of var ia t ion in to the err or te rm and m ake them corr e la ted with th e
inst r um ents , mak ing the m odel speci fica t ion ser iously b iased . The second te s t exam ines th e
hypothesis of no aut ocorr e la t ion in th e e rror te rm. The pr esence of f irs t -order
autocorr e la t ion in th e d i ffer enced r esidua ls does not imply th a t th e est imates a r e
inconsisten t , but th e pr esence of second -order au tocor r elat ion implies th at th e est imates ar e
inconsisten t . Th e second -order au tocorr elat ion test is r eported.7
Stage 2 Invest igat ing the de term inants o f th e persis tence o f prof i t
At stage 2, the deter min ant s of variat ions in th e coun tr y -level persist ence of pr ofi t
coefficients ar e exam ined. The est imated j obtained from s tag e 1 ar e th e dependen t variable
in a regr ession in which coun tr y and bank ing s ector char ac ter is t ics a re the independent
variables. The ind ependent var iables are g rou ped into th e following four categories:
macroeconomic and f inan c ia l system; r egula tory; inst i tu t iona l and exte rna l govern ance ;
and mar ke t s t ru c ture and compet it ion .
The macr oecon omic an d fina ncial system contr ol variables ar e in flation, r eal GDP
g row th a n d m ark et capitalization . In f la t ion i s t he ann ua l pe rcen tage chan ge in th e coun t ry
GDP deflator . Th e ban kin g ma r k et is less l ik ely to be competi t ive wh en i t is su bject to high
inflat ion, as t h e pr ices of fin an cial services, such as in terest r ates, are less infor ma tive
(Claessens an d Laeven, 2004), an d will in tu rn exacerbate credit mar k et frict ions (Boyd et
al , 2001). Ang elini an d Cetorelli (2003) find evidence of a neg ative relat ions hip between
6 The Sarg an t est is another tes t of over identifying condit ions as discussed in Ar el lano an d Bond (1991). This tes t ,
however, r equires homoscedast ic error s for consis tency, which is not necessari ly the case in our larg e cross-
coun try d ata sam ple. Hence we report Han sen s tat is t ics only.
7 The fai lure to pass ei ther of the tes ts in some cases sug gests th at th ese est imates ar e not completely free of
s tat is t ical problems. However, instead of using different methodologies for different coun tr ies , a consis tentmeth odology is adopted so that th e resul ts ar e compar able.
8/18/2019 Bbswp10002 Ok
10/43
10
inf la t ion and compet i t ion , sug gest ing an associa t ion be tween h igh inf la t ion and the
per s is ten ce of p r of it .
A posit ive r elat ionsh ip is expected between real GDP grow th and business
opportu ni t ies for banks. The incr eased business opportun i t ies may h e lp bank s to susta in
th eir pr ofi ts. Therefor e an association m igh t be expected between g rowth in GDP and t he
per s is t en ce of p r of it . On th e ot h er h an d, th e ava il abi li ty of bu sin es s op por tu n it ies m ig h t
lead to an inten sification of competi t ion , in wh ich case a neg ative relat ion sh ip would be
expected between GDP gr owth a nd th e pers istence of profi t .
Mar k et capit al iz at ion , the va lue of l is ted shar es on th e domest ic s tock ma rk e t
divided by GDP, is an indicator of th e size of the stock m ar ket r elat ive to th e econ omy.
Claessens an d Laeven (2004) ar g ue for a posit ive association between m ar k et capital ization
and compet i t ion , on t he basis tha t a h igh ly deve loped n on-bank ing f inan c ia l sector m ay
apply compet i t ive pressur e upon the ban king sector . Therefore a nega t ive re la t ionsh ip i s
expected between ma rk et capitalization an d th e persisten ce of pr ofi t .
The regu latory contr ol variables include pr oxies for r estr ict ions on bank activity,
capita l requir ements , and lega l barr iers to ent ry , obta ined f rom th e Bar th e t a l . (2004,
2006, 2008) data base. A ct iv it y res t r ic t ion is a measu r e of a bank ‟s abili ty to eng ag e in t h r ee
fee-based activit ies, secur it ies, insu r an ce an d r eal estate , in ad dit ion to inter est -based
ba n k in g act iv it y .8 This indicator is scaled from 0 to 12, wher e a hig h er value ind icates mor e
onerous r est r ict ions. Rest r ic t ions on bank ac t iv ity mig ht impr ove the sa fety an d soun dness
of the bank ing system, by minimiz ing opportu ni t ies for bank s to take on excessive r i sk ,
elimina ting some conflicts of inter est , and simp lifying ban k su pervision. However,
restr ict ions m igh t also r educe or el imin ate oppor tu nit ies for bank s to realize econom ies of
sca le and scope , or to man age th e var iabil ity in earn in gs across pr oduct l ines .
Total fraction of entry denied i s th e nu mber of ent r y appl ica t ions denied as a
fraction of th e nu mber of applications r eceived from d omestic an d foreign bank s over t he
per iod 2001 to 2005 . Th is m ea su r e r ef lect s on e a spec t of ba r r ie r s to en t r y , an d a p os it ive
8 Bank s in both developed an d developing coun tries h ave increased th e proport ion of their operat ing income
obtained from n on-interest sour ces in r ecent years . While part of the increase in n on-interest income is due to
diversificat ion int o lines of business such as in vestment banking , ventur e capital and insur ance under wr i t ing ,
gr owth in fee-paying and commission-paying services l ink ed to t radi t ional retai l banking services has also been
significant . Most evidence sugg ests that the sh ift towards non-interest income has n ot improved r isk-adjusted
retur ns (St iroh and Rum ble, 2006; Goddard et al . , 2008; St iroh, 2009).
8/18/2019 Bbswp10002 Ok
11/43
11
relat ionsh ip is expected between th is measu re an d the persist ence of pr ofit . Dom es t ic
f rac t ion of en t ry den ied and for eig n f rac t ion of en t ry den ied are u sed as a l tern a t ive barr ier
to ent ry measur es in separa te es t imat ions .
Capital requirem ents i s defined as th e minimu m capita l to asse t ra t io required by
govern ment r egula tory agenc ies . High er capital requirem ents incr ease the size of ent r y
ba r r ie r s , t h u s in s u la t in g in cu m ben t ba n k s fr om com pet it ion , an d in cr ea sin g ba n k pr of it
per s is ten ce.
Fin an cial f reedom i s a measu re of bank ing secur i ty as wel l as a m easur e of
independen ce from g overn m ent contr ol. Accordin g to th e Ind ex of Economic Freedom
pu bl is h ed in 2009 , th is m ea su r e is det er m in ed by th e ex ten t of g over n m en t r eg u la t ion of
f inan c ia l se rv ices , the amou nt of s ta te in te rvent ion in f inan c ia l inst i tu t ions, en t ry barr iers
to the finan cial services area, an d g overn m ent in fluen ce on th e al location of cr edit . The
high er th e score , the less th e govern ment inf luence and hence the m ore f inan c ia l f reedom .
An association is exp ected between th e quali ty of th e coun tr y‟s inst i tut ional and
externa l g overnan ce f ramework and pers is tence of prof it . Govern ance ru les an d pr ac t ice
differ between coun tr ies, and ar e depend ent on th e level of economic an d finan cial
developm ent (La Porta et a l . , 1998, 2002; Ch ha ochh ar ia and Laeven , 2009). Four
inst i tu t iona l var iables a r e us ed , cover ing proper ty r ig hts , inst i tu t iona l deve lopment ,
economic freedom, an d GDP per capi ta . Protec t ion of proper ty r ight s i s an importan t pr e-
requ isi te for a well-fun ctionin g finan cial system . Th e p rop er ty r ig h t s covariate , measured
on a sca le of 1 to 5 wi th h igh er scores indica t ing s t r onger pro tect ion , i s 6 m i n u s the
averag e proper ty r ight s f reedom index repor ted by the Heri tage Found at ion over th e per iod
1997-2007. The K K Z in dex (Kaufm an et a l ., 2008 ) measur es the level of ins t i tut ional
deve lopment , based on s ix d im ensions of governa nce , wi th h igh er scores re f lec t ing a m ore
advanced level of development. Th e aver ag e valu e for th e period 1998-2007 is u sed. The
degr ee of freedom enjoyed by busin esses and in dividu als from g overn m ent int erference is
measu red by the econom ic freedom covariate . This indicator is 6 m i n u s th e averag e
economic fr eedom ind ex r eported by th e Heritag e Fou nd ation over t he period 1997 -2007,
measu red on a sca le of 1 to 5 wi th h igh er scores re f lect ing gr ea ter f reedom. As a s tan dard
8/18/2019 Bbswp10002 Ok
12/43
12
measu re of econom ic developm ent, logar i thm of per capit a GDP is fou nd by Claessens an d
Laeven (2004) to influen ce bank ing sector per forma nce.
HHI , the Herf indah l-Hirschm an Index, is u sed as a concentr a t ion m easure for each
coun try‟s bankin g sector .9
Accord ing to eith er th e collusion h ypoth esis or th e efficiency
hy pothesis, a posit ive relat ion sh ip is exp ected between concentr ation an d th e persistence of
p r of it (Ber g er , 1995; God dar d et a l. , 2001). Accor d in g to th e „qu ie t li fe ‟ hypot h es is ,
however, a n egative relat ionsh ip between concentr ation an d persist ence of pr ofi t mig ht be
expected i f the man agers of la r ge bank s choose to set t le for a quie t l ife inst ead of pur suin g
pr of it s (H ick s , 1935 ; Ber g er an d Han n an , 1998). Th e Pan za r an d Ros se (1 987) H-s tat is t ic is
us ed as an al ter n ative cond u ct-based measu r e of competi t ive cond it ions.10
Typically, th e
empir ical li teratu r e on ban k competi t ion th at u ses th e H-stat ist ic report s evidence of
mon opolist ic competi t ion (Car bo et a l . 2009 ; Goddar d an d Wilson, 2009).11
4. Data and results
9 For an industr y that consis ts of a s ing le monopoly producer, HH 1 . A monopolis t has a mar ket sh are of s
1 1 .
For an indu stry w ith N firm s, the maximu m possible value of the Herfindahl – Hirschman index i s HH 1 , and the
minimu m possible value is H H 1 /N. The max imum value of HH 1 occur s when the s ize dis tr ibut ion of the N
firm s is h igh ly skewed. In the most extr eme case, one dominant f i rm h as a mar ket shar e only fract ional ly smaller
than 1 , and N 1 very sm all f i rms each has a m ark et shar e only fract ional ly lar ger th an zero.
10 Under monopoly condit ions H
8/18/2019 Bbswp10002 Ok
13/43
13
Finan cial accou nt s data (un con solidated) for th e period 1997 to 2007 for 1 1,634
ba n k s fr om 65 cou n t r ie s a r e obta in ed fr om Ba n k scope. Th e sam ple select ion cr it er ia a r e a s
follows. Any coun tr y with five or m ore com m ercial bank s (domestic or foreig n) is selected
for t h e firs t-stage an alysis. Th e star t-year, 1997, is const r ained by the availabili ty of data
on Bank scope. To eliminat e outl iers, obser vations for wh ich the pr ofit ra te was mor e than
thr ee s tand ard devia t ions f r om th e sample mean of each coun try a re de leted . Coun tr ies for
which j i s es t imated a t the f irs t s tag e using d a ta for fewer t han 15 bank s a r e el imina t ed
from the second-stage an a lysis .
The da ta for th e macroeconomic and r egula tory var iables u sed in th e second -stag e
an alysis ar e obtained from th e following sour ces. Macroeconomic data (inflat ion , GDP per
capita , r eal GDP g rowth , stock m ar ket capital ization) ar e obtained from th e World Bank ‟s
World Development In dicator s. The reg ulator y var iables (activity r estrict ions , capital
requir ements) and th e domest ic and fore ign ent r y var iables a re obta in ed from Barth e t a l .
(2008). The fin an cial fr eedom , econom ic freedom and p r operty r igh ts var iables a r e obtained
from th e Economic Freedom Index of the Heritag e Fou nd ation; the KKZ index is obtained
from Ka ufm an n et a l . (2008 ). Full details of sour ces an d variable definit ions a re l isted in
Table 1.
Table 1 Data sources and variable definiti ons
The fina l sample for the f i rs t -s tage ana lysis i s an un balan ced pan e l wi th 46 ,959
ba n k -yea r obser va t ion s (11 ,6 34 ba n k s). Th e p r of it r a t e m ea su r e is r et u r n on aver ag e eq u it y
(ROE), measu r ed by th e rat io of net income after tax to total equity. 1 2
Taxes ar e dedu cted,
beca u se if ef fect ive tax r a tes d if fe r be tween cou n tr ie s , en t r y an d ex it s h ou ld be d r iven by
after-tax pr ofi ts. The profi t rate of bank i located in coun tr y j a t year t , i,t
, is nor malized by
deduct ing the m ean pr ofi t ra te of a ll bank s loca ted in coun try j in year t .
12 The empirical analysis is also conducted us ing a retur n on averag e assets (ROA) profi tabi li ty measur e. To savespace these est imations are n ot reported, but th ey are avai lable from the au thors u pon request .
8/18/2019 Bbswp10002 Ok
14/43
14
Table 2 an d 5 collectively r eports d escriptive stat ist ics for r etur n on equity.
Summ ary da ta a r e tabula ted for a l l count r ies , based on averag es for th e per iod 1997-2007
in Table 5. Aver ag e profi tabil i ty varies sig nifican tly between cou nt r ies. The lowest aver ag e
ROE are -9.10% (Ur ug ua y) and -4.29 (Tha iland). Th e high est ar e 25.37% (Venezuela) and
23.84% (Nigeria). Aver ag e profi tabil ity was h igh er in developing cou nt ries (10.87%) than
in developed cou nt ries (6.51%). By region (Table 2), Nor th American bank s r eported th e
hig hest pr ofi tabil i ty on averag e (12.76%). East Asian a nd P acific r egion bank s r eported the
lowest pr ofi tabil i ty on averag e (1.06%). Table 3 r eports d escriptive stat ist ics for th e
dependent var iable and covar iates used in th e second -stag e analysis, wh ile Table 4 pr esents
the correlat ion matrices.
Insert Tables 2 and 3 Descriptive s tatis tic for profit rates an d other variables
Insert Table 4 Correlat ion Matrices
Firs t -s tag e an aly sis : es t im at in g th e per sis t en ce of prof it
Table 5 r eports est im ates of the sh ort -r un pers istence of profi t coefficients by
country , j in (5). Coun tr ies for wh ich th e est imated per sistence of pr ofi t coefficien t l ies
outside th e ran g e (-1, 1) are omitt ed from Table 5.13
Insert table 5 P rofit persis tence parameters -- tw o step dynamic panel data model
In Table 5, th e est imated persist ence of profi t coefficien t is posit ive for 62 of the 65
coun tr ies, and n egative for th r ee cou nt ries (Costa Rica, Dominican Repu blic an d Sout h
Korea). The mean j for the ent i re sam ple i s 0 .43 , somewhat h igher than the m ean va lue for
s ix Eu ropean countr ies of 0 .26 r epor t ed by Goddard e t a l . (2004a). The resul ts su gg est tha t
the e limina t ion of abnormal pr ofi t s thr oug h compet it ion i s by no m eans instan tan eous.
13
Microeconomic theory pr edicts tha t th e dynam ic process of competi t ion wil l res tore pr ofi ts to a nor mal r etur n.Assuming tha t -1≤ ≥ 1 pr ofi ts will converge t o the equil ibrium ra te of retur n over t ime.
8/18/2019 Bbswp10002 Ok
15/43
15
Barr ie rs to ent r y an d exi t a r e suff ic ient ly h igh to enable bank s to r e ta in a s ign i ficant
por t ion of th ei r abn or m a l pr of it s fr om yea r to y ea r .
Table 5 r eports th e fr equen cy distribution of the est imat ed per sistence of profi t
coefficients, j. For 40 of th e 65 coun tr ies, th e est im ated
j l ies within th e int erval 0.2 to
0.6.14
Developed cou nt ries ha ve sl igh tly hig her pers istence of profi t (averag e j = 0.442)
than deve loping coun tr ies (averag e j = 0.426), but th ese differ ences are n ot sign ificant.
Insert table 6 frequency distribution of the persis tence coefficients
Table 6 also reports th e averag e est imated j for coun tr ies located in eigh t
inter nat ional „reg ions‟: Latin America and Car ibbean; East Asia an d th e Pacific; Nort h
America; West Eur ope; East E ur ope; Middle East an d Nort h Africa; Sub -Sahar an Africa;
an d Sou th Asia. There is con siderable r egional variat ion in th e averag e per sistence of
pr of it . Th e h ig h es t aver ag e j ar e obtained for N orth America (aver ag e = 0.680) and
Wester n E u rope (avera g e = 0.495). The lowest values are obtain ed for Eas t Asia and th e
Pacific (averag e = 0.288) an d Su b-Sahar an Africa (averag e = 0.296).
Second -stag e analysis: identi fy ing th e determ inant s of persisten ce of profi t
The second-stag e an alysis exa min es the deter min an ts of th e persist ence of pr ofi t .
Table 7 r eports th e est ima tion r esults. In colu mn 1, only th e macroeconom ic variables
(inflat ion, real GDP gr owth, stock m ar ket capital izat ion) are includ ed in th e est im ation. In
colum ns 2-7 , the r egula tory var iables (ent r y bar r ie rs , ac t iv ity r est r ict ions , barr ie rs to
domest ic bank s, and bar r ie rs to fore ign banks ) a re added, one a t a t ime. In colum ns 8 -11
the inst i tu t iona l and exte rn a l gover nan ce var iables (proper ty r igh ts , KKZ index, GDP per
capita an d economic freedom) ar e added. In column s 12 -13 th e indu st ry concentra t ion an d
comp eti t ion m easu res (HHI and H -stat ist ic) are added. Finally, in colum n 1 4 all covar iates
14 The definition of advanced and developing countries follows the IMF, International Financial Statistics (2008).
8/18/2019 Bbswp10002 Ok
16/43
16
ar e in clud ed in a m ultivar iate model of th e determ inan ts of per sistence of pr ofi t .
Windm eijer sma ll sam ple corr ection is used to obtain r eliable standa rd er ror s.15
Insert Table 7 Cros s country determinants of s hort run profits persis tence
Among th e macroecon omic var iables, inflat ion g enerally ha s a posit ive imp act on
th e per sistence of pr ofi t , a l thoug h on ly a few of the coefficient s ar e sign ificant. In colum n
4, ther e i s a s ign i ficant nega t ive re la t ions hip be tween s tock mar ke t capi ta l iza t ion an d t he
per s is ten ce of p r of it . Th is wou ld su g g es t th a t th e m or e h ig h ly develop ed is th e n on -ba n k
fina ncial sector, the gr eater is the comp eti t ive pr essu re on th e bank ing sector . However,
th is f ind ing is r eversed in colum n 1 4. Overall , th erefor e, th ere is l it t le eviden ce for an y
relationsh ip between finan cial sector developmen t an d th e per sistence of ban k p rofi t . The
coeff ic ients on GDP gr owth a re pr edominan t ly nega t ive , and s ig ni f icant in man y cases .
This is consis tent wi th th e a rg um ent tha t g rea te r business opportu ni t ies a fforded by
high er economic growth lower th e opportu ni ty cost of en te r in g th e bank ing sector , thereby
redu cing entr y barr iers. Subsequ ently, in creased competi t ion weakens th e abil i ty of
incumbent bank s to su sta in th e ir prof itab il ity .
The coefficients on th e thr ee lega l barr ier s to ent r y indicators in colum ns 2 , 6 and 7
are posit ive and s ign i ficant . This sug gests th a t successfu l bank s in mar ke ts th a t a re
pr ot ected by en t r y ba r r ie r s a r e m or e able to su s ta in th ei r p r of it abi li ty over t im e. Th e
remaining regulatory variables - activity restrict ions, capital regulation, and financial
freedom - seem to exert l it t le or n o influ ence on th e persistence of pr ofi t .
Overa l l, regu la t ion appears to impact on th e pers is tence of pr ofi t thr ough the
erect ion an d enforcement of lega l bar r ie rs to ent r y . High er ent r y barr iers tend to reduce
comp eti t ion, and incr ease profi tabil ity. The opposite is a lso tru e. For in cum ben t bank s,
however, other t ypes of reg ulation do not exert a ny discern able impa ct on competi t ion
measu red by the persisten ce of pr ofit .
15 The control variables are selected to avoid high correlation, and problems of multicollinearity are thereforeminimal .
8/18/2019 Bbswp10002 Ok
17/43
17
Among th e inst i tutiona l and extern al gover na n ce variables, the coefficient on th e
pr op er ty r ig h t s m ea su r e is n eg a t ive an d s ig n if ican t in colu m n 11. Th is s u g g es ts th a t th e
st ron ger i s th e pro tect ion of proper t y r ight s , the lower i s the pers is tence of pr ofi t . With
st ron ger pr oper ty r igh ts , new entr ants m ay fee l more secur e in cond uct ing busin ess , so the
inten sity of competi t ion incr eases. The coefficients on t h e oth er inst i tu tional variables ar e
neg ative in colu mn s 9 and 10 , but th ese coefficient s ar e not sig nifican t.
There is no r elat ionsh ip between th e HHI index a nd t h e pers istence of profi t ,
according to the est imat ion r epor ted in colum n 12 . Differen ces between t h e objectives of
ba n k s m ay r en d er th e r el a t ion sh ip be tween in d u s t r y s t r u ct u r e an d con cen tr a t ion t en u ou s.
For exam ple, some banks migh t sacr i f ice potent ia l prof it s in order to r educe r i sk by
mak ing more secur e investments; or a l tern a t ive ly , some bank s m ight s imply enjoy a quie t
l ife afforded by a lack of competi t ive discipline (Berg er an d Han na n, 19 9 8; Goddar d et a l . ,
2004b). Similarly, ther e is n o relat ions hip between t h e Rosse-Pan zar H-stat ist ic and th e
per s is ten ce of p r of it , accor din g to th e es t im at ion r ep or ted in colu m n 13 . In colu m n 14 ,
however, wh ere both th e HHI index a nd t he H-stat ist ic are included, the coefficients on both
ar e sig nifican t. The relat ions hip between t h e HHI an d th e per sistence of profi t is n egative,
in accordan ce with th e „collus ion‟ hy pothesis of an association between concentr ation an d
entr y barr iers . Th e r elat ionsh ip between th e H-stat ist ic an d th e pers istence of pr ofi t is
neg ative. This findin g is also cons istent with an association between th e int ensity of
comp eti t ion an d th e persistence of pr ofi t .
In th e mu ltivar iate model r epor ted in colum n 14 , the coefficients on th e total
fraction of entr y denied, pr operty r igh ts an d GDP gr owth variables become insig nifican t. It
appears th a t th e mark e t s t ru c ture an d compet i t ion var iables tend to domina t e the ent r y
ba r r ie r s , in st it u t ion a l an d m acr oecon om ic va r ia bles in ex pla in in g th e va r ia t ion in th e
per s is ten ce of p r of it .
5. Conclus ion
Over th e past thr ee decades, bankin g m ar ke ts a r ound t he world have exp er ienced
funda menta l regu la tory an d s t ru c tura l chang e , wi th profoun d impl ica t ions for th e
8/18/2019 Bbswp10002 Ok
18/43
18
in tensi ty of bank compet i t ion . Using a la r ge c ross-coun try dynam ic pane l da ta se t , th is
paper r ep or t s es t im ates of th e in t en sit y of com pet it ion , ba sed on a dyn am ic an aly sis of th e
per s is ten ce of ba n k p r of it over t im e. In con t r ast wit h th e es sen t ia lly s t a t ic s t r u ct u r e -
conduct-performance (SCP) and new empirical industrial organization (NEIO) frameworks,
which domina te th e ana lysis of compet it ion in th e empir ica l ban king l ite ra tu re , th is paper
adopts a dyn am ic view of bank comp eti t ion. The present app roach is consistent with th e
view, a r t icu la ted in the contestable mark e ts l ite ra tu re , tha t thr ea tened or potent ia l en t ry i s
as importan t as ac tua l en t ry in de te rm ining the in tensi ty of compet i t ion .
A two-stag e ana lysis of the pers is tence of profi t i s repor ted usin g da ta on banks
from 65 coun tr ies over th e per iod 1997 to 2007. At th e fi rs t s t age , a f irs t -order
au toreg r essive model is used to est imate th e sh ort -ru n per sistence (speed of adju stm ent) of
ba n k p r of it . Th e speed of con ver g en ce va r ies be tween cou n t r ie s , wit h ba n k s in develop ed
coun tr ies exhib it ing h igh er pers is tence of prof it on averag e than those in developing
countries. Although competi t ion exerts some discipline on bank profi tabil i ty, adjustment
towards equi l ibr ium af ter an abnorm al profi t has been earn ed is by no means
instan taneous. Bar r ie rs t o ent ry a nd exi t a r e suff ic ient ly h igh to enable banks to re ta in a
signi f icant por t ion of th e ir abnorm al prof it s f r om year to year .
At the second sta g e, th e determ inan ts of th e pers istence of pr ofi t are exam ined. The
estimated pers istence coefficients fr om th e first sta g e ar e used as th e dependen t variable in
regr essions wi th covar ia tes th a t m easur e macroeconomic condi t ions, bank regu la t ion , the
inst i tu t iona l environm ent and exte rn a l governan ce, and compet it ion . The resul ts su gg est
tha t the pers is tence of profi t i s s t rong er in m ark e ts tha t a r e she l tered f rom compet it ion by
legally imposed and enforced bar rier s to entr y.
From an ant i t r ust , reg ula tory and s upervisory persp ect ive , the approach tha t i s
adopted in t h is paper su gg ests tha t policy r ecommend at ions based on pr ofi t s preva il ing in
the bank ing sector a t any poin t in t ime ar e inadvisable , as cur rent pr ofi tabi li ty i s
essential ly a tr an sitory ph enomen on. A dyn am ic model of profitabili ty provides an
indication of th e efficiency of competi t ion in forcing th e adjustm ent or converg ence of
pr of it s (a bove or be low th e n or m ) tow ar ds th ei r lon g -r u n eq u il ibr iu m va lu es . Th is ass is t s
the r egula tor in d is t ing uishin g between cases in which a compet i t ive equ i libr ium is l ike ly
8/18/2019 Bbswp10002 Ok
19/43
19
to be achieved r apid ly wi thout in te rvent ion , an d cases in wh ich regu la tory in te rvent ion
ma y be requ ired in order to achieve a comp eti t ive out come.
8/18/2019 Bbswp10002 Ok
20/43
20
References
Agostino, M., Leonida, L. , and Trivieri , F. , (2005) P r ofi ts per sistence and owners hip:
evidence from t he Ital ian ban kin g s ector, A ppli ed Econ om ics , 37, 161 5-21.
Ang elini , P. an d N. Cetorelli , (2003) Effects of regu latory r eform on competi t ion in bank ing ,
Jou rn al of Mon ey , Cred it an d Ban k in g , 35, 663-684 .
Arellan o, M., Bond , S., (1991) Some Tests of Specification for Pa n el Data: Mont e Car lo
Eviden ce and an Application to Em ployment Equ ations, R ev iew of Econ om ic S tu die s 58,
277 – 297.
Arellan o, M. an d Bover, O. (1995) An other look at th e ins tr u men tal var iable est ima tion of
error-components models, Jou rn al of Econ om et rics , 68, 29 -51.
Athan asoglou, P. P. , Brissim is, S.N. and Delis, M.D. (2008) Ban k -specific, Indu str y-specific
an d ma croeconomic Determ inan ts of Bank Pr ofi tabili ty, Jou rn al of In ter n at ion al Fin an cial
Mar k et s In s t it u t ion s an d Mon ey , 18, 121-136.
Barth , J .R., Capr io, G., Levin e, R. , (2004) Bank r egu lation an d su pervision: wh at work s
best ? Jou r n al of Fin an cial In t er m ed iat ion , 13, 205 -248.
Bar th , J.R., Capr io, G., Levin e, R (200 6) R e-t h in k in g Ban k R eg u la t ion : T il l A n g els Gover n .
Cam bridg e: Cam bridg e Universi ty Pr ess.
Barth , J.R. , Capr io, G., Levine, R (2008) Ban k Reg ulations Ar e Ch an gin g: For Better or
Worse? Comp arative Econom ic Studies , 50, 537-63
Baum ol, W, Pan zar, J an d Will ig, R, (1982) Contestable m arke ts an d th e theory o f
indus tr ial s t ruc ture . New York : Har cou rt Br ace J ovan ovich.
http://www.ingentaconnect.com/content/pal/ces;jsessionid=66ik5uno3qses.victoriahttp://www.ingentaconnect.com/content/pal/ces;jsessionid=66ik5uno3qses.victoriahttp://www.ingentaconnect.com/content/pal/ces;jsessionid=66ik5uno3qses.victoriahttp://www.ingentaconnect.com/content/pal/ces;jsessionid=66ik5uno3qses.victoria
8/18/2019 Bbswp10002 Ok
21/43
21
Bektas , E. (2007) The pers is tence of prof it s in th e Tur kish bank ing system, A ppli ed
Econ om ics Let t er s , 14, 187-190.
Berg er , A.N. , (1995) The prof it -s t ru c ture r e la t ions hip in bankin g: tes ts of mar ke t power
and eff ic ient s t ru c tur e hypotheses . Jou rn al of Mon ey , Cred it an d Ban k in g 27, 404 -431.
Berg er, A.N. (2007) Obstacles to a Global Bank ing System: Old Eu r ope versu s “N ew
Europe , Jou rn al of Ban k in g an d Fin an ce , 31, 1955-1976.
Berg er, A.N. an d Han na n, T.H. (1998) The efficien cy cost of ma rk et power in th e bank ing
indu str y: a test of th e qu iet l ife an d r elated hyp otheses, R ev iew of Econ om ics an d S tat is t ic s ,
80, 454 – 65.
Berg er, A.N., Kashya p, A.K. and Scalise, J.M. (1995) The tr an sform ation of the US
Banking indu st r y : what a long s t r ang e t r ip it ‟s been, B rook in g s Paper s on Econ om ic
A ct iv it y , 2, 54-219.
Berg er, A.N., Bonim e, S.D., Covitz, D.M., Han cock , D., (2000) Wh y ar e ban k p r ofits s o
per s is ten t? Th e r oles of pr od u ct m ar k et com pet it ion , in fo r m at ion op acity an d r eg ion al
macroeconomic shocks. Jou rn al of B an k in g an d Fin an ce 24, 1203-1235.
Berg er, A.N., Demerg u c-Ku nt , A., Levin e, R. , Haubr ich, J.C. , (2004 ) Ban k concent ra tion
an d comp eti t ion. Jou rn al of Mon ey , Cred it an d Ban k in g 36, 433 -451.
Black, S .E. and St rah an, P .E. (2002) Entr epreneur ship a nd bank credit ava i labil ity , Jou rn al
of Finance, 57, 2807-2833.
Blun dell, R and Bond S (1998) Init ia l Condit ions an d Moment R estrict ions in Dyn am ic
Pan el Data Models. Jou rn al of E con om et ric s , 87, 115-143.
8/18/2019 Bbswp10002 Ok
22/43
22
Bona ccorsi Di Pat ta , E. a nd Dell‟Aricca, G. (2004) Bank comp etit ion a nd firm creation,
Jou rn al of Mon ey Cred it an d Ban k in g , 36, 225-252.
Boyd J .H., Levin e R., and Sm ith B.D. (2001 ), Th e Impact of Inflat ion on Finan cial Sector
Performance , Jou rn al of Mon et ar y E con om ics 47, 221 -248.
Bra t land J . (2004) Contestable mar ke t Theory as a Regu la tory Fra mework: An Aust r ian
Postmortem, Quarterly Journal o f Aus tr ian E conom ics, 7, 3-28.
Carbo, S. and Fernan dez, F . (2007) The Determ inants of Bank Marg ins in Eu ropean
Bank ing , Jou rn al of Ban k in g an d Fin an ce, 31 2043-2063.
Car bo, S. Hu mp h rey, D. Maudos, J. and Molyneu x, P (2009) Cross -coun tr y comp ar isons of
compet i t ion and pr icing power in Eu ropean bank ing, Jou rn al of In t er n at ion al Mon ey an d
Fin an ce, 28 115 – 134
Cetorelli , N. and Str ah an , P.E. (2006) Fina nce as a bar rier to entr y: bank comp eti t ion an d
indus t ry s t r uc tu re in loca l US bankin g m ark e ts , Jou rn al of Fin an ce , LXI, 437 - 461 .
Ch am berlain , E. (1933). The th eory of monopolist ic comp eti t ion. Cam bridg e MA: Har vard
Universi ty Press.
Ch ha ochh ar ia , V. an d Laeven, L. (2009) Corpor ate govern an ce nor m s and p r actices,
Jou rn al of Fin an cial In t er m ed iat ion , for thcoming .
Clark e, G., Cull , R. , Mar tinez Per ia , M.S. and Sanch ez, S.M. (2003) F oreign Bank En tr y:
Ex perience, Implications for Developing Economies, and Agend a for Fu r th er Research ,
World Bank R esearch Observer , 18, 25 -59.
8/18/2019 Bbswp10002 Ok
23/43
23
Claessens, S., Demirg u c-Ku nt , A. and Hu izing a, H. (1998) How Does Foreign En tr y affect
the domest ic bank ing sector? World Bank Work ing Paper No: 1918
Claessens, S. , an d Laeven, L. , (2004) What dr ives ban k comp eti t ion? Some intern ational
evidence. Jou rn al of Mon ey , Cred it an d Ban k in g 36, 563 -83.
Dick, A.A. and Han na n, T.H. (2009) Competi t ion and an ti tr us t in bank ing , in Berg er, A.N.,
Molyneux, P. and Wilson, J.O.S. (eds.) Oxford Handbook o f Bank ing . Oxfor d: Oxford
Universi ty Press.
De Youn g, R. , Hun ter, W., an d Udell , G., (2004) The past , pr esent an d pr obable futu re for
commu ni ty banks , Jou rn al of Fin an cial Ser v ices R es earch , 25, 85-133.
Demirg uc-Kun t , A., and Huiz ing a , H. , (1998) Determin ants of commerc ia l bank in te rest
ma r gin s and pr ofi tabil i ty: some intern ational eviden ce, World Bank Econom ic Rev iew , 13,
379-408.
Flamin i, V., McDona ld, C.A., an d Sch um acher , L.B. (2009 ) The Determin ants of
Comm ercial Bank Pr ofitabili ty in Sub-Sah ar an Africa, IM F W or k in g Paper N o. 09 /15
Gerosk i, P.A. (1990 ) Modelling p ers isten t pr ofitability, in Mueller, D.C. (ed.) The Dyn amics
of Com pany Profi ts: A n In tern ation al Comp arison. Cam bridg e: Cam bridg e Univer si ty Pr ess.
Gerosk i, P.A. (1990 ) Modelling p ers isten t pr ofitability, in Mueller, D.C. (ed.) The Dyn amics
of Com pany Profi ts: A n In tern ation al Comp arison. Cam bridg e: Cam bridg e Univer si ty Pr ess.
Ger oski, P.A. and J acquem in, A. (1988) The pers istence of pr ofi ts: a Eur opean compar ison,
Econ om ic Jou rn al , 98, 37 5-89.
http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=935549http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=907776http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=78767http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1356442http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1356442http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1356442http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1356442http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=78767http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=907776http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=935549
8/18/2019 Bbswp10002 Ok
24/43
24
Glen, J . , Lee, K. and Sin gh , A. (2001) P ersisten ce of pr ofi tabil ity a nd comp eti t ion in
emerging mar ke ts : a t ime ser ies ana lysis , Econ om ic Let ter s , 72, 247 -53.
Glen, J. , Lee, K. and Sing h , A. (2003) Corpor ate pr ofi tabil i ty and th e dyn am ics of
compet i t ion in em erging mar ke ts : a t ime ser ies an a lysis , Econ om ic Jou rn al , 113, F465-
F484.
Goddar d, J .A. and Wilson, J .O.S., (1996) Per sistence of profi t s for UK manu factu r ing an d
service sector fi r ms , Service Industries Journal , 16, 105 – 17.
Goddar d, J . and Wilson, J .O.S. , (1999) Per sistence of pr ofit : a n ew empirical interp r etat ion ,
In t er n at ion al Jou rn al of In du s tr ia l Or g an iz at ion , 17, 663 -87.
Goddar d, J . , Molyn eux , P. , and Wilson, J .O.S. , (2001 ) Eu rop ean ban k in g : E f f ic ien cy ,
techn ology and g rowth . J ohn Wiley, Chichester , UK.
Goddar d, J . , Molyn eux , P. , an d Wilson, J .O.S., (2004a) The pr ofitabili ty of Eur opean ban k s:
A cross-sectional and dy na mic panel an alysis. Man ch es ter Sch ool , 72, 363-381.
Goddar d, J. , Molyneu x , P. , an d Wilson, J .O.S., (2004 b) Dynam ics of gr owth a nd
pr of it abi li ty in ba n k in g . Jou rn al of Mon ey , Cred it an d Ban k in g 36, 1069-1090.
Goddar d, J. , Molyn eux , P., Wilson , J.O.S. an d Tavakoli, M. (2007) Eu r opean Bank ing : an
overview, Jou rn al of Ban k in g an d Fin an ce , 31, 1911-1936.
Goddard, J., McKillop, D.G. and Wilson, J.O.S. (2008) The diversification and financial
per for m an ce of US Cr ed it Un ion s‟, Jou rn al of Ban k in g an d Fin an ce, 32, 1836-49
Goddar d, J. an d Wilson, J .O.S. , (2009) Competi t ion in bank ing : a disequil ibriu m a ppr oach,
Jou rn al of B an k in g & Fin an ce , for thcoming.
8/18/2019 Bbswp10002 Ok
25/43
25
Han sen, L. (1982) Lar ge sam ple pr opert ies of generalized meth od of m oment s est imator s,
Econ om et rica, 50, 1029 – 1054.
Heri tage Found at ion (2009) In dex of Econ om ic Fr eedom , 2009 , Met h od olog y for th e 10
Econ om ic Fr eedom s.
Hick s, J. (1935). Ann ua l sur vey of econom ic th eory : The theory of mon opoly , Econ om et rica
3, 1-20.
IMF (2008) In t er n at ion al Fin an cial S tat is t ics . Washin gton : In te rn a t iona l Monetary Fu nd.
J ones, K.D. and Critchfield, T. (2005) Consolidation in th e US bank ing ind us tr y: is th e
„lon g, st r an g e trip‟ about to end ? FDIC Ban k in g R ev iew , 17, 31-61.
Kau fman n D., Kr aay, A. and Mas tr u zzi , M. (2005) Govern an ce Matter s IV: Govern an ce
Indicators for 1996-2004. World Bank Pol icy R esearch Work ing Paper , Number 3630.
Kau fman n D., Kraay , A. and Mast r u zzi , M. (2008 ) Govern an ce Matters VII: Governa nce
Indicators for 1996-2007. World Bank Pol icy R esearch Work ing Paper , Number 4654.
Kn app, M., Gar t , A., and Chau dh r y, M. (2006) The impact of pers istence and m ean
reversion of bank pr ofi tabi li ty on post -merg er performan ce in th e bank ing indu st r y ,
Jou rn al of B an k in g an d Fin an ce , 30, 3503 -3517.
La Port a , R. , Lopez-de-Silan es, F. , Shleifer , A. and Vish ny , R. (1998) Law a nd Fina nce,
Jou rn al of Poli t ical Econ om y , 106, 1113 – 55.
La Por ta , R. , Lopez-de-Silan es, F. , Shleifer , A. an d Vish ny , R. (2002 ) Investor Pr otection
and Corporate Valuation, Jou rn al of Fin an ce , 57, 1147 – 70.
8/18/2019 Bbswp10002 Ok
26/43
26
Levine R. , (2003) Denying Fore ign Bank En try Impl ica t ions for Bank In te rest Marg ins ,
Central Bank of Chile W ork ing Papers , No. 222.
Levonian , M.E. (1993) The per sistence of bank pr ofi ts: what t he stock m ar ket im plies”,
Wor kin g Paper s in Applied Econom ic Theory, No. 93-15. San F ra n cisco, Federal Reserve
Bank of San Fr anc isco
Lipczynsk i, J. , Wilson, J .O.S. and Goddar d, J. (2009 ) In du st r ia l Or g an iz at ion : Com pet it ion ,
Strategy, Policy . Har low: FT Pr entice Hall.
Maray um a, N. an d Odagir i , H. (2002) Does the persisten ce of pr ofi ts pers ist? A stu dy of
compan y pr ofi t s in J apan, 1964 -1997, In ter n at io n al Jou rn al of In du s t r ia l Or g an iz at ion , 20,
1513-1533.
Maudos, J . and Fern andez de Guevara , J . (2004) Fac tor s a ffect ing the in te rest m arg in in
the bank ing sectors of the Eur opean Union, Jou rn al of Ban k in g an d Fin an ce , 28, 225 9-81.
Maudos, J . and F ernan dez de Guevara , J . (2007) The cost of mar ke t power in bank ing:
social welfar e ver su s cost inefficiency, Jou rn al of Ban k in g an d Fin an ce , 31, 2103-2126.
McGahan , A.M. and P orter , M.E., (1999) The per sistence of shocks to pr ofi tabil i ty. R ev iew
of Econom ics and Statist ics 81, 143-153.
Mueller, D.C., (1977) The per sistence of profi ts above the n orm . Econ om ica 44, 369-380.
Mueller, D.C., (198 6) Pr of it s in th e lon g ru n . Cam bridg e Univers i ty Pr ess, Cam bridg e, UK.
Pan zar, J .C. and Rosse, J.N. (1987 ). Testin g for m onopoly equ il ibriu m. Jou rn al of
In du s tr ia l Econ om ics , 35, 443 -456.
8/18/2019 Bbswp10002 Ok
27/43
27
Roland, K.P. , (1997) Pr ofi t persisten ce in lar ge U.S. ban k h olding comp an ies: an empir ical
investigation, Econ om ics W or k in g Paper , 97 -2 . Of f ice of th e Com ptrol ler of Cu rren cy .
Roodm an , D. (2006) How to do xt abond2: an int rodu ction to “difference” and “syst em”
GMM in Stata , Center for Global Developm ent W ork ing Paper Num ber 103 .
Roodman , D. (2008) A note on the theme of too man y inst r um ents , Cent er f or Global
Develop m en t W or k in g Paper Nu mber 125 .
Shaffer, S. , (2004). Pattern s of competi t ion in bank ing , Jou rn al of Econ om ics an d Bu sin es s ,
56, 287 -313.
Steph an, A. and Tsapin , A. (2008) Persis tence and Determin ant s of Fi rm Prof it in
Emerg ing Marke t s , CESIS W ork ing Paper 151.
Stiroh, K. an d Str ah an , P.E. , (2003) Competi t ive dyn amics of deregu lation: evidence for US
ba n k in g , Jou rn al of Mon ey , Cred it an d B an k in g , 35, 801-28.
Stiroh, K.J. (2009) Diversification in ba n kin g, in Berg er, A., Molyn eux , P. and Wilson,
J .O.S. (eds.) Oxford Handbook o f Bank ing . Oxford: Oxford Universi ty Press, forthcoming.
Stiroh, K.J . and Ru mble, A. (2006) The dar k side of diversification: The case of US finan cial
holding companies, Jou rn al of Ban k in g an d Fin an ce , 30, 2131-2432.
Vander Venn et, R., (2002) Cost an d pr ofit efficiency of finan cial con glomer ates an d
un iversa l bank s in Eu rope , Jou rn al of Mon ey , Cred it an d Ban k in g , 34, 254 -82.
Waring, G. (1996) Industry Differences in the Persistence of Firm-Specific Returns,
A m er ican Econ om ic R ev iew , 86, 1253 -65.
8/18/2019 Bbswp10002 Ok
28/43
28
8/18/2019 Bbswp10002 Ok
29/43
29
Table 1 Variable definitions and sources
Variable name Description and source
Dependent Var iable
Return on average equity (ROE)2 Total net income after tax divided by average equity, as reported by Bankscope.
Bank market structure and competiti on 2 :
Herfindahl Index A measure of the degree of concentration in the banking industry, calculated by the sum of the squares of each bank's marketshares of the country of the year.
H-statisticA measure of the degree of competitiveness of the industry, with less than 0 being a monopoly, less than 1 while larger than
0 being monopolistic competition, and 1 being perfect competition .
Macroeconomic variabl es 5 :
Inflation rate Rate of inflation, calculated by log difference of GDP deflator.
GDP growth The growth in GDP.
Market Capitalization Total domestic stock market capitalization divided by GDP
Regulatory restricti ons:
Activity Restrictions1
A measure of a bank's ability to engage in the businesses of securities underwriting, insurance, and real, and of the
regulatory restrictiveness of banks to own shares in non-financial firms (sum of Survey of Bank Regulation and Supervision
questions 4.1 through 4.4).Greater values signify more restrictions.
Financial freedom3 An indicator of financial freedom (ranging from 0 to 100). Calculated as 100 minus the financial freedom index of the
Heritage Foundation. Higher values signify less freedom.
8/18/2019 Bbswp10002 Ok
30/43
30
Capital requirements1 The minimum capital to asset ratio required by the government (Based on question 3.1 of Survey of Bank Regulation and
Supervision).
Fraction of entry denied1 A measure of the number of entry applications denied as a fraction of the number of applications received from domestic
and foreign entities (ratio of Survey of Bank Regulation and Supervision questions 1.9.1 and 1.10.1 to 1.9 and 1.10).
Domestic fraction of entry denied1
A measure of the number of entry applications denied as a fraction of the number of applications received from domestic
entities (ratio of Survey of Bank Regulation and Supervision questions 1.9.1 to 1.9).
Foreign fraction of entry denied1 A measure of the number of entry applications denied as a fraction of the number of applications received from foreign
entities (ratio of Survey of Bank Regulation and Supervision questions 1.10.1 to 1.10).
I nstitutional environment :
Economic freedom3 An indicator of economic freedom (ranging from 0 to 100). Greater values signify more freedom.
Property rights3 An indicator of the protection of private property rights (ranging from 0 to 100). Greater values signify better protection of
property rights.
KKZ Institution index4
An indicator of the quality of institutional development in the country. Calculated as the average of six indicators: voice and
accountability, political stability, government effectiveness, regulatory quality, rule of law, and control of corruption.
GDP per capita
5
Logarithm of GDP per capita expressed in original values.
Data Sources:
1 Barth et al. (2008). Data available at: www.worldbank.org/research/projects/bank_regulation.htm
2 Calculated based on Fitch IBCA's Bankscope Database
3 Economic Freedom Index of the Heritage Foundation
http://www.worldbank.org/research/projects/bank_regulation.htmhttp://www.worldbank.org/research/projects/bank_regulation.htmhttp://www.worldbank.org/research/projects/bank_regulation.htmhttp://www.worldbank.org/research/projects/bank_regulation.htm
8/18/2019 Bbswp10002 Ok
31/43
31
4 Kaufmann, Kraay, Zoido-Lobaton (2008)
5 World Development Indicators, World Bank
Table 2 Summary statistics for ROE by region and development.
Region Number of countries ROE
Latin America and Caribbean 16 10.07
East Asia & the Pacific 11 1.06
North America 1 12.76
West Europe 15 7.02
East Europe 11 11.47
Middle East & North Africa 2 7.10
Sub-Saharan Africa 6 13.22
South Asia 3 15.93
Advanced Countries 19 6.51
Developed Countries 46 10.87
Overall 65 10.10
8/18/2019 Bbswp10002 Ok
32/43
32
Table 3 Summary statistics for the persistence of profit, and covariates
MeanStandard
DeviationNumber of countries
Persistence of profit, j 0.429 0.243 64
Inflation rate 0.078 0.129 62
Market capitalization 0.546 0.643 59
GDP growth 0.040 0.018 62
Herfindahl-Hirschman Index 0.177 0.135 60
Rosse-Panzar H-statistic 0.553 0.255 62
Activity restrictions 2.554 0.650 56
Capital requirements 0.089 0.014 55
Financial freedom 46.393 15.568 61
Total fraction of entry denied 0.070 0.135 48
Domestic fraction of entry denied 0.093 0.194 31
Foreign fraction of entry denied 0.057 0.131 42
Economic freedom index 63.452 9.000 61
KKZ Index 0.388 0.910 63
Property rights Index 59.061 23.035 61
GDP per capita 8.572 1.463 61
8/18/2019 Bbswp10002 Ok
33/43
33
Table 4 Correlation Matrices for the explanatory variables.
InflationMarket
capitalization
GDP
growthHHI
H-
statistic
Activity
restrictions
Captial
requirements
Financial
freedom
Totalentry
denied
Domesticentry
denied
Foreignentry
denied
Economic
freedomKKX
GDP per
capita
Inflation 1.00
Marketcapitalization
-0.28 1.00
0.03
GDP growth 0.05 -0.19 1.00
0.68 0.15
HHI -0.01 -0.14 0.11 1.00
0.92 0.30 0.41
H-statistic 0.08 -0.06 -0.16 0.06 1.00
0.54 0.64 0.23 0.66
Activity
restrictions0.09 -0.41 0.13 0.11 -0.07 1.00
0.50 0.00 0.36 0.43 0.63
Captialrequirements
0.54 -0.39 0.18 0.04 -0.08 0.03 1.00
0.00 0.00 0.19 0.80 0.55 0.84
Financial
freedom 0.23 -0.51 0.30 0.15 0.00 0.43 0.32 1.00
0.08 0.00 0.02 0.26 0.97 0.00 0.02
Total entry
denied-0.02 -0.15 0.25 0.10 0.05 0.16 0.06 0.31 1.00
0.88 0.31 0.09 0.53 0.73 0.28 0.68 0.04
8/18/2019 Bbswp10002 Ok
34/43
34
Domesticentry denied
0.21 -0.04 0.36 0.18 0.01 0.18 0.05 0.51 0.90 1.00
0.27 0.84 0.05 0.35 0.95 0.34 0.77 0.00 0.00
Foreign
entry denied-0.06 -0.19 0.30
-
0.06-0.02 0.13 0.16 0.17 0.81 0.45 1.00
0.73 0.24 0.06 0.71 0.89 0.44 0.34 0.31 0.00 0.02
Economic
freedom-0.38 0.71 -0.33
-
0.140.12 -0.34 -0.54 -0.80 -0.33 -0.42 -0.32 1.00
0.00 0.00 0.01 0.30 0.36 0.01 0.00 0.00 0.02 0.02 0.04
KKX -0.28 0.59 -0.38-
0.120.01 -0.46 -0.46 -0.67 -0.36 -0.38 -0.36 0.79 1.00
0.03 0.00 0.00 0.37 0.93 0.00 0.00 0.00 0.01 0.04 0.02 0.00
GDP percapita
-0.25 0.56 -0.47-
0.200.08 -0.49 -0.44 -0.67 -0.42 -0.42 -0.51 0.75 0.91 1.00
0.05 0.00 0.00 0.13 0.52 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.00
Notes: p value of each pair of correlation coefficients is reported underneath each correlation coefficient.
8/18/2019 Bbswp10002 Ok
35/43
35
Table 5 First-stage estimation results: Short-run persistence of profit estimates and profit rates (ROE) by country
CountryCountry
typeRegion
No.of
banks
No.
of
obs
ROE
Mean
ROE
Std.
Dev.
Lambda
Lambda
Std.
Err.
Hansenp AR(2)
Argentina DC Latin America and Caribbean 94 543 -3.29 37.14 0.164 0.008 0.302 0.546
Australia AC East Asia & the Pacific 26 87 11.52 21.86 0.255 0.085 0.477
Austria AC West Europe 248 1361 7.10 8.08 0.477 0.039 0.570 0.107
Bahamas DC Latin America and Caribbean 16 89 19.83 17.67 0.353 0.098 0.988 0.294
Bangladesh DC South Asia 31 198 17.94 17.03 0.284 0.034 0.192 0.811
Belgium AC West Europe 62 344 9.80 19.07 0.539 0.004 0.522 0.547
Bosnia and Herzegovina DC East Europe 23 124 6.28 10.09 0.525 0.033 0.474 0.313
Brazil DC Latin America and Caribbean 151 861 14.14 23.34 0.311 0.021 0.102 0.818
Bulgaria DC East Europe 22 132 13.15 16.33 0.392 0.008 0.444 0.220
Cayman Islands DC Latin America and Caribbean 16 82 11.29 11.48 0.613 0.046 0.999 0.387
Chile DC Latin America and Caribbean 34 197 9.26 10.76 0.400 0.005 0.474 0.451
China DC East Asia & the Pacific 70 345 12.78 10.08 0.570 0.025 0.823 0.727
Colombia DC Latin America and Caribbean 27 169 9.39 30.87 0.324 0.003 0.957 0.449
Costa Rica DC Latin America and Caribbean 45 208 13.14 7.14 0.887 0.016 0.866 0.832
Croatia DC East Europe 41 226 8.03 12.92 -0.061 0.002 0.601 0.230
Cyprus AC East Europe 15 71 16.17 28.65 0.508 0.011 0.990 0.136
Czech Republic DC East Europe 24 130 9.53 18.78 0.201 0.003 0.998 0.102
Denmark AC West Europe 104 636 10.70 7.92 0.488 0.010 0.105 0.990
Dominican Republic DC Sub-Saharan Africa 33 175 9.92 28.90 -0.034 0.002 0.863 0.102
Ecuador DC Latin America and Caribbean 33 218 12.21 12.79 0.169 0.009 0.301 0.286
Egypt, Arab Rep. DC Sub-Saharan Africa 30 181 8.49 18.43 0.517 0.002 0.828 0.135
France AC West Europe 343 2022 9.17 19.08 0.446 0.024 0.331 0.031
Germany AC West Europe 2123 12178 5.41 5.57 0.178 0.017 0.000 0.002
Guatemala DC Latin America and Caribbean 32 205 12.14 14.46 0.808 0.008 0.803 0.614
Honduras DC Latin America and Caribbean 20 122 11.27 8.90 0.723 0.016 0.599 0.866
8/18/2019 Bbswp10002 Ok
36/43
8/18/2019 Bbswp10002 Ok
37/43
37
Switzerland AC West Europe 442 2392 8.18 7.51 0.725 0.032 0.116 0.861
Taiwan, China DC East Asia & the Pacific 46 301 -0.80 17.44 0.325 0.022 0.356 0.101
Tanzania DC Sub-Saharan Africa 15 52 19.99 26.02 0.254 0.156 0.077
Thailand DC East Asia & the Pacific 21 135 -4.29 55.57 0.967 0.002 0.403 0.259
Tunisia DC Middle East & North Africa 15 94 6.95 7.12 0.630 0.028 0.840 0.034
Ukraine DC East Europe 33 188 11.71 11.34 0.454 0.017 0.599 0.796
United Kingdom AC West Europe 145 692 9.13 13.47 0.684 0.017 0.455 0.198
Uruguay DC Latin America and Caribbean 44 195 -9.10 82.89 0.589 0.010 0.692 0.665
United States AC North America 738 4319 12.76 9.27 0.680 0.041 0.098 0.004Venezuela DC Latin America and Caribbean 49 268 25.37 18.63 0.391 0.025 0.358 0.593
Vietnam DC East Asia & the Pacific 22 128 10.38 14.59 0.580 0.010 0.479 0.271
Note: is estimated by applying the system GMM estimator to Equation (4), with the profit rates being the ROE‟s (r eturn-on-average equity) as reported byBankScope. „AC‟ represents advanced countries, while „DC” represents developing countries.. 'Hansenp' is the p-value for the Hansen test for the validity of
the over-identifying restrictions in the System GMM estimation. 'AR(2)' is the p-value for the 2nd-order autocorrelation in the residuals from the System GMMestimation.
8/18/2019 Bbswp10002 Ok
38/43
38
Table 6 Frequency distribution of estimated short-run persistence coefficients
a) By frequency interval
Frequency intervalNumber of countries
AC DC all %
8/18/2019 Bbswp10002 Ok
39/43
39
Table 7: Second-stage regression results, determinants of the persistence of profit
Macro-
economic
conditions
Macroeconomic conditions and bank regulation
(1) (2) (3) (4) (5) (6) (7)
Constant 0.858* 0.714** 1.115 1.294 0.783 1.041*** 0.763
(2.82) (2.44) (1.93) (1.77) (1.51) (3.97) (2.02)
Inflation rate 2.512** 1.881 2.305 -0.252 2.613 -3.387 1.756
(2.19) (1.44) (1.25) (-0.27) (1.25) (-1.59) (1.28)
Marketcapitalization
0.0137 0.0758 0.00322 -0.0836* 0.0298 -0.0517 0.0637
(0.19) (1.28) (0.04) (-3.01) (0.25) (-0.25) (0.98)
GDP growth -20.91** -21.35* -23.29 4.167 -21.05** -15.39 -21.55**
(-2.19) (-2.78) (-1.66) (0.75) (-2.34) (-1.85) (-2.33)
Total entry denied 1.871***
(3.75)
Activity
restrictions-0.0711
(-0.72)
Capitalrequirements
-11.34
(-1.09)
Financial freedom 0.00133
(0.15)
Domestic entry
denied0.964**
(2.47)
Foreign entrydenied
1.804**
(2.54)
8/18/2019 Bbswp10002 Ok
40/43
40
N 61 46 53 52 59 31 40
adj. R-sq 0.482 0.476 0.434 0.826 0.417 0.931 0.291
Note: Dependent variable is coefficients estimated from the first stage system GMM estimator. Since the dependent variable is anestimated parameter, all equations are weighted with the inverse of the standard errors of the dependent variable. The constant variable is
suppressed; however, the inverse of standard error of dependent variable is used instead. OLS with robust standard errors is applied in each
equation and t statistics of estimated coefficients are shown in parentheses.*,**,*** represents 10, 5, and 1 percent significant level,respectively
Institutional and External GovernanceMarket
StructureCompetition
ll
(8) (9) (10) (11) (12) (13) (14)
Constant 3.307 1.096* 2.632** 1.794* 0.663 1.355* -0.115
(1.91) (3.08) (2.62) (3.13) (1.87) (3.24) (-0.20)
Inflation rate 1.383 1.830** 1.892** 0.554 2.694** 1.627 0.560
(0.88) (2.06) (2.32) (0.45) (2.26) (1.53) (0.86)
Market
capitalization0.298 0.0842 0.0932 0.106 0.0260 -0.0366 -0.107*
(1.52) (1.28) (1.46) (1.85) (0.33) (-0.53) (-2.74)
GDP growth -25.89* -24.80** -27.17* -25.74* -19.22 -22.94** 2.181
(-3.05) (-2.46) (-2.72) (-3.07) (-1.98) (-2.64) (0.41)
Total entry denied -0.00136
(-0.00)
Economic freedom -0.0360
(-1.47)
8/18/2019 Bbswp10002 Ok
41/43
41
KKZ -0.211
(-1.60)
GDP per capita -0.170
(-1.99)
Property rights -0.0106** 0.0127
(-2.01) (2.45)
HHI 0.514 1.035***
(1.11) (3.64)
H-statistic -0.348 -1.100***
(-1.30) (-6.88)
N 59 60 61 59 61 59 44
adj. R-sq 0.526 0.541 0.593 0.554 0.496 0.565 0.795
Note: Dependent variable is coefficients estimated from the first stage system GMM estimator. Since the dependent variable is an estimated parameter, all equations are weighted with the inverse of the standard errors of the dependent variable. The constant variable is suppressed;however, the inverse of standard error of dependent variable is used instead. OLS with robust standard errors is applied in each equation and t
statistics of estimated coefficients are shown in parentheses.*,**,*** represents 10, 5, and 1 percent significant level, respectively
8/18/2019 Bbswp10002 Ok
42/43
42
8/18/2019 Bbswp10002 Ok
43/43