Risk of Credit and In­for­ma­tional Asym­me­try

Economic Challenger - - CONTENTS - - Ayadi Fatma

Ab­stract:

In the case of fi­nan­cial dis­tress, the bank­ing de­ci­sion is lim­ited to the alternative be­tween bank­ruptcy or fi­nan­cial sup­port. Gen­er­ally, when in­for­ma­tion is ac­ces­si­ble, the bank chooses the de­ci­sion which max­i­mizes its profit. In ad­di­tion, when in­for­ma­tion is pri­vate, the bank chooses the de­ci­sion to bank­ruptcy sanc­tion as the illiq­uid com­pany. On the con­trary, if the bank has an in­for­ma­tional ad­van­tage com­pared to the other cred­i­tors, it is able to ex­ploit it and in this case, the prob­a­bil­ity of the state­ment of af­fairs of an illiq­uid com­pany is then re­duced at least in the short term in the form of an abu­sive sup­port.

We pro­pose to check em­pir­i­cally the im­pact of the in­for­ma­tional prob­lems on the bank­ruptcy of the com­pa­nies start­ing from the Bel­gian ac­count­able data. A Pro­bit model is used. The en­doge­nous vari­able is the fact of hav­ing or not to file for bank­ruptcy, the ex­plana­tory vari­ables in­clude the re­veal­ing vari­ables of the fi­nan­cial, eco­nomic and le­gal sit­u­a­tion of the com­pa­nies, as well as of the proxy vari­ables of in­for­ma­tional asym­me­tries. The as­sump­tion of the abu­sive sup­port is not al­ways con­firmed. The re­sults on the con­trary ap­pear that the de­ci­sion to bank­ruptcy is all the more se­vere than in­for­ma­tion is pri­vate and hardly ac­ces­si­ble. Key words: credit risk- In­for­ma­tional Asym­me­try- Bank­ing De­ci­sion-

IN­TRO­DUC­TION

Since the Eight­ies, sev­eral the­o­ries in­sisted on the fact that the risk of credit con­sti­tutes the first in­ter­nal cause of bank­ing fail­ure. It is de­fined as "non per­for­mance" of the coun­ter­part gen­er­at­ing a prob­a­ble loss on the level of the bank.

More pre­cisely, it comes from the in­ca­pac­ity of the debtor to re­spect all or part of its obli­ga­tions to­wards the cred­i­tor. This last will un­dergo this non-pay­ment; it can de­ter­mine the ca­pac­ity to re­quire to file the bal­ance-sheet of the debtor. How­ever, it may find it ben­e­fi­cial to give a fi­nan­cial sup­port for its cus­tomer in hope of a fur­ther re­fund­ing.

In ad­di­tion, the de­ci­sion of credit put back on the ef­fec­tive treat­ment of in­for­ma­tion by the bank, which re­quires the need for an or­gan­i­sa­tional struc­ture adapted to the type of in­for­ma­tion. In­deed, it is in this con­text of the bank­ing re­la­tion that the asym­me­try of in­for­ma­tion be­tween the parts can cause a prob­lem of mo­ral risk lead­ing to an in­ef­fec­tive de­ci­sion of credit. Thus, it is al­ways nec­es­sary to fol­low con­trol by mech­a­nisms of ad­e­quate or­ga­ni­za­tion and gov­er­nance bank­ing. In ad­di­tion, there are other causes be­ing able to in­flu­ence the de­ci­sion of bank credit as in­sti­tu­tional and le­gal en­vi­ron­ment, reg­u­la­tion and pru­den­tial su­per­vi­sion of the banks, as well as the struc­ture and the dis­ci­pline of mar­ket. The lat­ter con­sti­tute pil­lars 1, 2 and 3 of the Basle II re­form. Th­ese pro­cesses sup­pose that the bank holds cap­i­tal funds in ad­e­quacy with its risk pro­file mak­ing it pos­si­ble to im­prove the com­mu­ni­ca­tion of in­for­ma­tion and to en­cour­age the ap­pli­ca­tion of the benef­i­cent bank­ing prac­tices.

Bank­ing in­dus­try is strongly de­pen­dant on in­for­ma­tion. The bank is al­ways con­fronted with the prob­lem of asym­me­try of in­for­ma­tion, and to solve it, it can ac­quire two types of in­for­ma­tion: Hard in­for­ma­tion, in ex­ter­nal, by the means of pub­lic in­for­ma­tion (Rat­ing, Score...) and of Soft in­for­ma­tion, in in­ter­nal, by the means of the re­la­tion of cus­tomers. That sup­poses two meth­ods of loans at­tri­bu­tion: the bank with the act ver­sus the bank of re­la­tion (Peter­son, 2004)

It is well-known in the lit­er­a­ture that the di­vi­sion of in­for­ma­tion and the re­spect of the in­ter­na­tional stan­dards im­prove the con­di­tions

of ac­cess to the credit by re­duc­ing the prob­lems of ad­verse se­lec­tion and mo­ral risk. The banks must be or­ga­nized to ap­pre­hend and treat the risk of credit, in par­tic­u­lar in­ter­na­tional statu­tory con­straints sup­port­ing on them (Basle II). The ob­ject of this ar­ti­cle is to pro­pose the prin­ci­pal the­o­ret­i­cal de­ter­mi­nants of the bank­ing de­ci­sion fac­ing the non-pay­ment of a debtor and to pro­pose there­after an em­pir­i­cal check­ing of their im­pact to bank­ruptcy. With this in­ten­tion, we will use the Pro­bit model to de­ter­mine the prob­a­bil­ity to bank­ruptcy through in par­tic­u­lar two sam­ples of healthy and fail­ing com­pa­nies be­long­ing to the Bel­gian sec­tor for pe­riod 20012004 and through a se­ries of in­di­ca­tors or vari­ables re­lated to the eco­nomic, fi­nan­cial and le­gal sit­u­a­tion of the com­pany and ac­count­able vari­ables able to re­duce the ex­is­tence of in­for­ma­tional asym­me­tries. We are in­spired partly by the em­pir­i­cal stud­ies where the in­for­ma­tional prob­lems seem like ex­oge­nous vari­ables, know­ing that it is dif­fi­cult to ap­pre­hend di­rectly the in­for­ma­tional prob­lems; the stud­ies con­sider vari­ables which make it in­di­rectly pos­si­ble to give an ac­count of it.

In the case of par­tic­i­pa­tion of sev­eral cred­i­tors in the com­pany in ad­di­tion to the bank, there is risk of ap­pear­ance of the con­flicts of in­ter­est and the prob­a­bil­ity of ‘to go into liq­ui­da­tion’ in­creases be­cause the fi­nan­cial sup­port falls on the bank (Bu­low and Shoven, 1978).

If on the other hand a coali­tion be­tween the share­hold­ers and the prin­ci­pal bank of the com­pany is suit­able to be trained to the detri­ment of bond cred­i­tors and the prob­a­bil­ity of a sup­port for the com­pany would be the most strong.

1- THE DATA

Ini­tially, we se­lect the Bel­gian com­pa­nies start­ing from the base pro­vided by World scope Global, then, we present re­veal­ing ac­count­able vari­ables of the ex­is­tence of the prob­lems of in­for­ma­tional asym­me­tries. Ac­cord­ing to some em­pir­i­cal stud­ies, we sup­pose that th­ese prob­lems seem ex­oge­nous vari­ables. Fi­nally we de­scribe, the ex­plana­tory vari­ables se­lected to de­scribe the eco­nomic, fi­nan­cial and le­gal sit­u­a­tion of the firm.

1.1 com­pa­nies:

The se­lected sam­ple in­cludes healthy com­pa­nies and fail­ing com­pa­nies. All are firms of the sec­tor of the in­ter­me­di­ate goods sub­ju­gated to the cor­po­ra­tion tax.

The de­ci­sion of set­ting in bank­ruptcy of the com­pany by the bank oc­curs at the time of the non-pay­ment. Thus, we con­sider for the fail­ing com­pa­nies the last as­sess­ment avail­able be­fore the bank­ruptcy, the data are pro­vided for 4 years (2001-2004).

The Year 2001 be­ing best rep­re­sented, it is that which we also re­tained for the non fail­ing com­pa­nies ( were not the sub­ject of le­gal pro­ce­dures).

We re­tain the as­sump­tion ac­cord­ing to which the fail­ing com­pany does not re­spect its fi­nan­cial en­gage­ments i.e. which had the fi­nan­cial ex­penses higher than the gross sur­plus of ex­ploita­tion.

50 fail­ing com­pa­nies and 120 non-fail­ing com­pa­nies were thus se­lected.

1.2 the choice of the vari­ables:

The in­for­ma­tional prob­lems are ap­pre­hended with dif­fi­culty, but some vari­ables in­di­rectly make it pos­si­ble to give of it an ac­count and they are gath­ered in the con­text of re­duc­tion of in­for­ma­tional asym­me­tries. - Vari­ables able to re­duce the ex­is­tence of the in­for­ma­tional prob­lems:

To re­duce the asym­me­tries ex­is­tence of in­for­ma­tion, one of the first vari­ables se­lected is the size of the firm or more pre­cisely the turnover or the to­tal num­bers. In­deed, more the com­pany is of sig­nif­i­cant size, more it is sub­jected to le­gal con­straints in term of pro­duc­tion of in­for­ma­tion. We thus re­tain the to­tal num­bers (NUM).

A sec­ond vari­able used is the du­ra­tion of the re­la­tion passed with the lender or the age of the com­pany. The re­la­tions of long term weaken asym­me­tries of in­for­ma­tion, re­duce the costs of mon­i­tor­ing and gen­er­ate there­after a re­lax­ation of the su­per­vi­sion of the com­pany by the bank.

Asym­me­tries of in­for­ma­tion are also re­duced ei­ther when the com­pany is with di­men­sions be­cause the lat­ter will be obliged to pro­duce re­li­able in­for­ma­tion or when the bank is

share­holder of the com­pany be­cause in this case it has ac­cess to the pri­vate in­for­ma­tion more eas­ily than as a sim­ple cred­i­tor.

This ta­ble checks off the var­i­ous re­veal­ing vari­ables of asym­me­tries of in­for­ma­tion:

The re­veal­ing vari­ables of growth op­por­tu­nity are the ex­pen­di­ture in re­search and devel­op­ment which rep­re­sent a source of fu­ture growth. We thus re­tain the re­la­tion­ship be­tween the ex­pen­di­ture in R&D and the turnover noted: (RD/ TUR)

Also, the in­tan­gi­ble cred­its rep­re­sent from ev­ery an­gle growth ap­pro­pri­ate­ness. In­deed if the ac­count­able as­sess­ment records them with their his­tor­i­cal value, on the con­trary the fi­nan­cial mar­ket, by its eval­u­a­tion of the firm, ac­counts for the an­tic­i­pated growth ap­pro­pri­ate­nesses. The pro­duc­tive in­vest­ment which is at the ori­gin of the growth op­por­tu­ni­ties; thus, we con­sider the ra­tio pro­duc­tive cap­i­tal ex­pen­di­ture on to­tal as­sets (PI/ASS) Some stud­ies try to iden­tify the risk of sub­sti­tu­tion of as­sets. It is all the more weak as the pro­duc­tive struc­ture is rigid, there­fore that the ra­tio tan­gi­ble as­sets im­mo­bi­lized on to­tal as­sets is high or that the ra­tio of the pro­duc­tive equip­ment on the en­gaged cap­i­tal is sig­nif­i­cant, we re­tain the ra­tio (PEQ/K), in­deed the risk of sub­sti­tu­tion of as­set is all the more sig­nif­i­cant since the com­pany car­ries out new in­vest­ments. The ra­tio debts to­wards the group and of part­ners on to­tal as­sets, (PAR/ASS) is used as proof of the con­fi­dence of part­ners within the con­text of in­for­ma­tional asym­me­tries. In­deed more this ra­tio is high; more the risk of sub­sti­tu­tion of as­sets is weak, be­cause the as­so­ci­ates would be the first de­spoiled in the case of ex­ces­sively risky in­vest­ment.

2- Vari­ables re­lat­ing to the eco­nomic, fi­nan­cial and le­gal sit­u­a­tion of the com­pa­nies:

Other ex­plana­tory vari­ables in­tro­duced into the model will be sup­posed to af­fect the bank­ing de­ci­sion. In the case of liq­ui­da­tion, the sold goods are the tan­gi­ble per­ma­nent as­sets and stocks. So, we re­tain the mar­ket value of the firm as first ex­plana­tory, it will be ap­pre­hended by the ra­tio (im­mo­bi­lized tan­gi­ble as­sets + to­tal stocks) / to­tal as­set noted (LIQ). The cri­sis of illiq­uid­ity is mea­sured by the mar­gin, be­tween the gross sur­plus of ex­ploita­tion and the fi­nan­cial ex­penses normed by the noted turnover, (LIQ/ TUR) A third vari­able will be re­tained in our em­pir­i­cal model to re­alise of the clear sit­u­a­tion of the com­pany un­der con­sid­er­a­tion by the to­tal as­sets ra­tio / to­tal debt (ASS/DEB). The debt po­si­tion is con­sisted of the bank­ing loans, of the cur­rent bank­ing con­tests, other fi­nan­cial debts and debts of ex­ploita­tion; there­fore, we re­tain the re­la­tion­ship be­tween the whole of the short-term debt and the whole of fi­nan­cial debt con­sid­ered by the bank­ing ra­tio debt of short­term/ to­tal debt (shtD/DEB)

This ta­ble presents a syn­the­sis of the used ra­tios. Ex­plana­tory vari­ables that they rep­re­sent and the eco­nomic phe­nom­e­non which they de­scribe.

Ac­cord­ing to the de­scrip­tive static of th­ese var­i­ous ex­oge­nous vari­ables, we could con­firm the the­o­ret­i­cal re­sults. In­deed, through the two un­der­pop­u­la­tions, the fail­ing com­pa­nies are char­ac­ter­ized on the av­er­age by a liq­uid­ity ra­tio (LIQ/TUR) and a lower net sit­u­a­tion (ASS/DEBT) whereas by com­par­isons with the healthy com­pa­nies, they are strongly fi­nanced by short­term debts (shtD/DEBT)

For the vari­ables able to re­duce the in­for­ma­tional prob­lems, only av­er­ages of the to­tal num­bers (NUM) and ra­tios loan of part­ners / to­tal as­sets (PAR/ASS) which are in con­form­ity with the the­ory. In­deed, more the com­pany is of sig­nif­i­cant size, more it is sub­jected to con­straints in term of pro­duc­tion of in­for­ma­tion.

In the same way for the ra­tio (PAR/ASS), more this ra­tio is high, plus the risk of sub­sti­tu­tion of as­sets is weak, be­cause the as­so­ci­ates would be the first stripped in the event of ex­ces­sively risky in­vest­ment.

The pro­duc­tive cap­i­tal ex­pen­di­ture re­ported in the to­tal as­sets (PI/ASS), are higher for the non-fail­ing com­pa­nies. In­deed they rep­re­sent a sig­nal of the hope of the fu­ture profit. How­ever, the mar­ket value (LIQ) is not sig­nif­i­cantly dif­fer­ent be­tween the two sub-pop­u­la­tions. 3- The method­ol­ogy of re­search and es­ti­mates:

The method­ol­ogy of re­search con­sists in test­ing two models: a ba­sic model in per­fect in­for­ma­tion and a to­tal model by us­ing vari­ables able to re­al­ize the ex­is­tence of asym­me­tries of in­for­ma­tion.

We seek to check if the ex­is­tence of asym­me­try of in­for­ma­tion has an in­flu­ence on the fact that a com­pany files for bank­ruptcy or not by

tak­ing ac­count of the two as­sump­tions in the case of fi­nan­cial dis­tress.

In­deed, the cred­i­tor who un­der­goes the non­pay­ment holds the ca­pac­ity to re­quire the liq­ui­da­tion of the debtor (as­sump­tion 1). But, it can have also in­ter­est to give a fi­nan­cial sup­port in the hope of a later re­fund­ing (as­sump­tion2). Ac­cord­ing to this study. The di­choto­mous vari­able Y will be fixed and a model Pro­bit is used.

By di­choto­mous model, we un­der­stand a sta­tis­ti­cal model in which the ex­plained vari­able can take only two meth­ods (di­choto­mous vari­able).

Then, it con­cerns gen­er­ally to ex­plain the un­ex­pected ar­rival or the non-un­ex­pected ar­rival of an event.

Among the most known ap­pli­ca­bil­ity, we quote that which con­sists of the mod­el­ling of the de­fault risks in a loan re­la­tion, or any other form of con­tract of ser­vice (tele­phone sub­scrip­tion con­tract, con­tract of as­sis­tance etc...). We con­sider for ex­am­ple a di­choto­mous vari­able tak­ing two meth­ods: "rup­ture of the con­tract" and "con­tin­u­a­tion of the con­tract ". They are the tech­niques of bases here, meth­ods of scor­ing largely used in the bank­ing sec­tor and in the telecom­mu­ni­ca­tions sec­tor. We pose, [l, N]:

Let us sup­pose that we have N ob­ser­va­tions y, what­ever i = 1..., N: of a coded di­choto­mous en­doge­nous vari­able yi = 1 or yi = 0 per con­ven­tion, when in par­al­lel the ob­ser­va­tions of K ex­oge­nous vari­ables are:

In this case, the sim­ple lin­ear model is writ­ten: Where in­di­cates a vec­tor of K un­known pa­ram­e­ters and where the dis­tur­bances.... are sup­posed to be in­de­pen­dently dis­trib­uted. Then, we can high­light sev­eral prob­lems in­volved in the use of this sim­ple lin­ear spec­i­fi­ca­tion to model our di­choto­mous vari­able. The vari­able yi is of qual­i­ta­tive type while the sum.

The di­choto­mous pro­bit model ad­mits for ex­plained vari­able, the prob­a­bil­ity of ap­pear­ance of this event con­di­tion­ally to the ex­oge­nous vari­ables. Thus, the fol­low­ing model is

Where the func­tion F (.) In­di­cates a distri­bu­tion func­tion. In the case of this model, the func­tion of distri­bu­tion F (.) cor­re­sponds to the func­tion of distri­bu­tion of the re­duced cen­tred nor­mal law

There­fore, for a given value of the vec­tor of ex­oge­nous and vec­tor of the pa­ram­e­ters we can de­fine the fol­low­ing model: For Yi is equal to 1 when a com­pany i is fail­ing and Yi is equal to 0 when com­pany i is non-fail­ing. X is the vec­tor of the ex­plana­tory vari­ables The Es­ti­mated Pro­bit model is writ­ten:

The er­rors ?i fol­low a re­duced cen­tred nor­mal law.

The vec­tor will be es­ti­mated by the method of the Max of like­li­hood.

The qual­ity of the ad­just­ment of the model is tested by the report of like­li­hoods.

Where F ex­presses the func­tion of distri­bu­tion of the re­duced cen­tred law.

This test con­sists in com­par­ing the

prob­a­bil­ity of the model con­sid­ered noted L with noted like­li­hood of the sim­ple model where the only ex­plana­tory vari­able con­sid­ered is the con­stant. The null as­sump­tion HO cor­re­sponds to the nul­lity of the whole of the pa­ram­e­ters. Un­der HO, the statis­tics S = 2(ln L -ln fol­low a law of chi-two of de­gree of free­dom

K (a num­ber of ex­plana­tory vari­ables ex­cept con­stant, are the di­men­sion of X mi­nus 1). The sig­nif­i­cance of each es­ti­mated pa­ram­e­ter βj (j=1,…, K-1) is also tested thanks to the report of like­li­hoods. The null as­sump­tion HO is the nul­lity of the pa­ram­e­ter ?j. The like­li­hood of the es­ti­mated model is com­pared with the like­li­hood of the con­strained model es­ti­mated un­der Un­der HO, the statis­tics fol­low chi-two of de­gree of free­dom 1.

In the case of the un­var­ied di­choto­mous model, the con­struc­tion of like­li­hood is ex­tremely sim­ple. In­deed, with the Yi event = 1 is as­so­ci­ated the like­li­hood and in event Yi=0 cor­re­sponds the like­li­hood . . This makes it pos­si­ble to re­gard the ac­tual val­ues Yi as the achieve­ments of a bi­no­mial process with a like­li­hood of

By dis­tin­guish­ing the ob­ser­va­tions Yi = 1 those for which we have Yi = 0, the log-like­li­hood can be writ­ten in the form: The es­ti­ma­tor of the max­i­mum of like­li­hood of the pa­ram­e­ters... is ob­tained by max­i­miz­ing ei­ther the func­tion of like­li­hood or the func­tion of log like­li­hood log

4-re­sults and con­clu­sions:

Es­ti­ma­tion of the model of ref­er­ence (per­fect in­for­ma­tion): From this model, we want to check that in the case of sym­me­try of in­for­ma­tion, the bank tends to sup­port its com­pany cus­tomer.

This ta­ble rep­re­sents for each ex­oge­nous vari­able J the es­ti­mate of the pa­ram­e­ter The num­ber be­tween brack­ets is the as­so­ci­ated prob­a­bil­ity. This model is over­all sig­nif­i­cant; each found sign con­firms the the­o­ret­i­cal re­sults. The sup­port of a com­pany is thus strongly cor­re­lated for the qual­ity of its eco­nomic and fi­nan­cial sit­u­a­tion. This sup­port which is gen­er­ally abu­sive. How­ever, such a con­clu­sion is to be mod­er­ated: If the sup­port is based on the re­li­a­bil­ity of the in­for­ma­tion held by the bank, then the ac­count­able as­sess­ments eas­ily avail­able al­ways do not ac­count for the true health of the com­pany? Es­ti­ma­tion of the model 2(in asym­me­try of in­for­ma­tion: in­tro­duc­tion of the in­for­ma­tional vari­ables).

To in­tro­duce th­ese vari­ables within the model, we will carry out an as­cend­ing se­lec­tion We carry out an as­cend­ing se­lec­tion (the high­est report of like­li­hood). We se­lect ini­tially the ex­plana­tory vari­able most cor­re­lated with the de­pen­dent vari­able. Then, we se­lect, among those which re­main, the ex­plana­tory vari­able whose par­tial cor­re­la­tion is high­est (by keep­ing con­stant the vari­ables al­ready se­lected). And so on as long as there re­main vari­ables can­di­dates whose par­tial co­ef­fi­cient of cor­re­la­tion is sig­nif­i­cant. We re­tained only the vari­ables from which the co­ef­fi­cient is sig­nif­i­cantly dif­fer­ent from zero with the thresh­old 10%. Ac­cord­ing to this as­sump­tion, one elim­i­nated from the model the co­ef­fi­cients as­so­ci­ated with the ex­pen­di­ture in re­search and devel­op­ment re­ported to the turnover and the share the pro­duc­tive equip­ment in F to­tal as­sets since they are not sig­nif­i­cantly dif­fer­ent from zero with the thresh­old of 0.1.

Only to­tal num­ber NUM, the pro­duc­tive cap­i­tal ex­pen­di­ture brought back in the to­tal as­sets.PI/ ASSand the share of the loans granted by the as­so­ci­ates in the whole of the to­tal as­sets.PAR/ASSare in­te­grated into model 1 to con­sti­tute model 2.

Model 2 is writ­ten then:

From this model, we em­pir­i­cally will check the im­pact of each in­for­ma­tional vari­able on the set­ting in bank­ruptcy. This ta­ble rep­re­sents for each ex­oge­nous vari­able, the es­ti­mate of the pa­ram­e­ter the num­ber be­tween brack­ets is the as­so­ci­ated prob­a­bil­ity:

This tested model is sig­nif­i­cant. We want more­over test the as­sump­tion HO ac­cord­ing to which the to­tal model is not sig­nif­i­cantly more ex­plana­tory than the ba­sic model.

Fol­low­ing the cal­cu­la­tion of the report of like­li­hood S=2(lnL the like­li­hood of model 2 and L is the like­li­hood of model 1.

The co­ef­fi­cients of the ex­oge­nous vari­ables re­lat­ing to model 1 re­main sig­nif­i­cant, with the awaited signs. The vari­ables in­tro­duced into model 2 have a neg­a­tive im­pact on the prob­a­bil­ity of set­ting in bank­ruptcy. This re­sult con­firms the idea that the set­ting in liq­ui­da­tion by the bank is in­flu­enced by the ex­is­tence of the prob­lems of asym­me­tries of in­for­ma­tion.

The pro­duc­tive in­vest­ment re­ported in the to­tal as­sets in­ter­venes neg­a­tively in the prob­a­bil­ity of set­ting in liq­ui­da­tion. In­deed, the growth ap­pro­pri­ate­nesses en­cour­age the bank to sup­port the com­pany.

Ac­cord­ing to th­ese re­sults, the de­ci­sion of set­ting in bank­ruptcy can be soft­ened only when asym­me­tries are re­duced.

As a con­clu­sion and through this method­ol­ogy of re­search, two as­sump­tions are taken into ac­count: the as­sump­tion of the sup­port of the com­pany which is strongly cor­re­lated with the qual­ity of its eco­nomic and fi­nan­cial sit­u­a­tion, in­deed, if pri­vate in­for­ma­tion would be held by the bank, all the ac­count­able data are eas­ily avail­able, the bank can de­cide in the case of fi­nan­cial dis­tress to sup­port its cus­tomer in the hope of a later re­fund­ing. The as­sump­tion of set­ting in bank­ruptcy which is all the more se­vere as pri­vate in­for­ma­tion is dif­fi­cult to reach, in this case, the bank in case of doubt and in fear of un­dergo the loss, the set­ting in au­to­matic bank­ruptcy of its cus­tomer de­cides. The es­ti­mated model can be more rel­e­vant by in­te­grat­ing sev­eral el­e­ments:

More de­tail by cred­i­tor of the au­tho­rized cred­its, this will be able to con­firm the the­o­ret­i­cal re­sults.

To in­crease the sam­ple of the com­pa­nies (healthy and fail­ing) while vary­ing the branches of in­dus­try this then sup­poses to in­te­grate a new vari­able in the model which is the vari­able Sec­tor.

To an­a­lyze the be­hav­iour of the bank over sev­eral pe­ri­ods and long-term this could im­prove the ex­pla­na­tion of the set­ting in bank­ruptcy of the com­pany

BIB­LI­OG­RA­PHY :

Bé­dué, Lévy, 1997, " re­la­tion ban­queen­treprise et coût du crédit ", re­vue d'économie fi­nan­cièren°39-1997. Bell,J et D Pain,2000," lead­ing in­di­ca­tor models of bank­ing crises_ a crit­i­cal re­view", fi­nan­cial sta­bil­ity re­view, banque d'an­gleterre, n°9, ar­ti­cle3,pp,113-129, décem­bre. Berger, An and Udell, GF, 1992," Some ev­i­dence on the em­pir­i­cal sig­nif­i­cance of credit ra­tioning," the jour­nal of po­lit­i­cal econ­omy,vol100,n°5, Oc­to­ber,1047-1077. Berger et udell, 1995,"Re­la­tion­ship lend­ing and lines of credit in small firm fi­nance", jour­nal of busi­ness, vol.68,n°3,pp351-377. Berger et udell, 2002," small busi­ness credit avail­abil­ity and re­la­tion­ship lend­ing: the im­por­tance of bank or­gan­i­sa­tion­nal struc­ture",eco­nomic jour­nal,112,pp34-53. Berk­son J. ( 1944), " Ap­pli­ca­tion of the Lo­gis­tique Func­tion to Bio-As­say ", JASA, 39, 357-365. Berk­son J. (1951), "Why I pre­fer Logit to Pro­bit", Bio­met­rics, 7, 327-339. Berlin M., Loeys J. (1988), " Bond Covenants and Del­e­gated Mon­i­tor­ing ", the Jour­nal of Fi­nance, vol. 43, p. 397-412. Berlin M, ET Mester L, 1992, " debt covenants and rene­go­ci­a­tion", jour­nal of fi­nan­cial in­ter­me­di­a­tion, vol.2, p95-133.

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.