应用财务比率预警财务危机

外文翻译原文T i tl e: Fi na ncia l R a ti os a nd the Proba bi l i sti c Pre di cti on o f B a nk ru ptcyM a te ria l S ou rc e: http: / / w w w .j stor. org /p ss /2 4 9 0 3 9 5 A u th or: J a m e s O hl son 1 I n t r od u c t ionT his pa per pre sents som e e m p i ri ca l re su l ts o f a stu d y pre di cti ng c orpo ra te fa i lu re a s e v i de nc ed by the e v e nt of ba nk ru ptcy . T h ere ha v e bee n a fa i r nu m ber o f prev i ou s stu di e s i n thi s fi e l d of re sea rch; the m ore nota b l e pu bl i she d c ontribu ti ons a re B ea v er ( 1 9 6 6 ; 1 9 6 8 a ; 1 9 6 8 b) , A l tm a n ( 1 9 6 8 ; 1 9 73 ) a nd so on. T w o u npu bli she d pa pers by W hi te a nd T u rnbu l i ( 1 9 7 5 a ; 1 9 7 5 b) a nd a pa pe r by S a ntom ero a nd V inso ( 1 9 7 7 ) a re of pa rti cu l a r intere st a s the y a ppea r to be the fi r st stu d i e s w hi ch lo g i ca l l y a nd sy ste m a ti ca l l y d e ve lop proba bil i sti c e sti m a te s of fa i lu re . T he pre sent s tu d y i s si m i l a r to the l a tter stu di e s, i n tha t the m ethodolog y i s one of m a x i m u m l i k e l i hoo d e sti m a tio n of the so- ca l l e d co nditio na l l o g i t m od e l .T he da ta se t u se d i n thi s stu dy i s from the se v enti e s ( 1 9 7 0 -7 6 ) . I k no w of only thre e co rpora te fa i lu re re sea rc h stu di e s w hi c h ha v e e x a m ine d da ta from th i s peri od. One i s a l im i ted stu d y by A l tm a n a nd M c G ou g h ( 1 9 7 4 ) in w hi c h o nl y fa i l ed fi rm s w ere dra w n from th e peri od 1 9 7 0 -7 3 a nd on l y o ne ty pe of cla ssi fica ti on error ( m i scla s si fica tion of fa i l e d fi rm s) w a s a na l y z e d. M oy er ( 1 9 7 7 ) co nsi d ered the peri od 1 9 6 5 -7 5 , bu t the sa m p l e of ba nk ru pt fi rm s w a s u nu su a l l y sm a l l ( tw enty - se v en fi rm s) . T he third stu d y , b y A l tm a n, H a l d e m a n, a n d Na ra y a na n ( 1 9 7 7 ) , w hi c h “u pd a te s“ the ori g i na l A l tm a n ( 1 9 6 8 ) stu dy , ba s i ca l l y co nsi d ers da ta from the period 1 9 6 9 to 1 9 7 5 . T heir sa m pl e w a s ba se d on fi fty -thre e fa i l e d fi rm s a nd a b ou t th e sa m e nu m ber of nonfa i l e d fi rm s . In c ontra st, m y stu d y rel i e s on observ a ti ons from 1 0 5 ba nk ru pt fi rm s a nd 2 ,0 5 8 n on ba nk ru pt fi rm s. A l thou g h the oth er thre e stu d i e s di ffer fro m the pre se nt one so fa r a s m ethod olog y a nd obj ec ti ve s a re c onc erned , i t i s, ne v e rth el e ss , intere sti n g a nd u se fu l to c o m pa re th e i r re su l ts w i th those pre se nted i n thi s pa pe r.A nother d i sti ng u i shi n g fea tu re of th e pre se nt stu dy w hi ch I shou ld stre s s i s tha t, contra ry to a l m ost a l l pre v i ou s s tu di e s, the da ta for fa i l e d fi rm s w ere not d eri v e d from M ood y 's M a nu a l . T he da ta w ere obta i ne d in s tea d from 1 0 - K fi na nc i a l sta te m en ts a s reporte d a t th e ti m e . T hi s proc e du re ha s one i m porta nt a dv a nta g e: there ports indi ca te a t w ha t poi nt i n ti m e the y w ere rel ea sed to the pu b l i c , a nd one ca n th ere fore che ck w h ether the c om pa ny en tere d ba nk ru ptc y prior to or a fter the da te of rel ea se . Prev iou s stu di e s ha v e not e x p l i citl y c onsid ere d thi s ti m i ng i s su e.2S om e C o m m e n t s R e g a r d in g M e th od o log y a n d D a t a C o l l e c t ionT he fu nda m e nta l e sti m a tion probl e m ca n be re du c e d s i m pl y to the fol low i ng sta te m ent: g i v en tha t a fi rm bel ong s to som e pre sp e ci fie d popu l a ti on, w ha t i s the proba bil i ty tha t the fi rm fa i l s w i thi n so m e pre spe ci fi e d ti m e peri od? N o a s su m ptions ha v e to be m a de reg a rdi n g pri or proba bili ti e s of ba nk ru ptc y a nd /or the di s tribu ti on of pre di ctors. T he se a re the m a j or a d v a nta g e s . T he sta ti sti ca l si g ni fi ca nc e of the di fferent pred i c tors a re obta i n ed from a sy m ptoti c ( l a rg e sa m pl e) the ory .C l ea rl y , m u c h ca n b e g a i ned by i m prov i ng the da ta ba se. T he e v a lu a ti on of the predi c tiv e cla s si f i ca tion pow e r of a m ode l sh ou l d b e m ore rea l i sti c , a nd, m ore im porta n t here, the sa m e shou ld a pply for s ta nda rd te s ts of s ta ti sti ca l si g n i fica nce . T his i s not to su g g e st tha t i t i s i m porta nt to ha v e “ su per a c cu ra te“ da ta for pu rposes of d e v e lopi ng ( a s oppose d to e v a lu a ting ) a di scri m ina tory d e v i ce . It m ig ht w el l be tha t th e pre di ctiv e qu a l i ty o f a ny m od el i s rea sona bl y robu st a cross a v a riety of da ta g a theri ng a nd e sti m a ting proce du re s.3C o l l e c t ion of F i n a n c i a l S t a