Full text of Review (Federal Reserve Bank of Dallas) : July 1991
The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
fEDERAL RESERVE RESERVE BANK BANK OF Of DALLAS DALLAS FEDERAL July July 1991 1991 • • conormc conOllllC eVlew eVleW I Misleadillg Indicatol's? Illdica/01~'/ Illis!eading IIsillg the /be COlnjJosile COllljJosile USilZg Leac/illg hldicaLoJ's /Ildiw/ors Leading /0 Predict Predic/ C:vclical C)'c/iwl to TUI'IIillg Poill/s 7~tl'1l ing PoinIs Evan F. F Koenig ilnd Jnd EVJn Kenneth M. Emery II/Mul'Jl La Pl'ojilabililJ Pmji/flbili/y 4 Relul'Jl 10 . . : 'llJe '[7)(! Peljorll7(lIIce Pt!lfOJ'/17ance 0/ (?/ Eler(,ll/1l Di.,/ric/ Flel'ellllJ Dz\/nd C(Yl771J? Jl'cial BallMs Ban!{..\' COllilllercial KpVlI'! Kevil; J. J Yeats This publication was digitized and made available by the Federal Reserve Bank of Dallas' Historical Library (FedHistory@dal.frb.org) Economic Review Federal Reserve Bank ofOa/las Flobert O. MeTflr, Jr. -~ ""'<-Ton., J. S.IY'ggio ""'I'u"-_ _.- a.MQ.MMlpMc. H.rvty Rosenblum _ .se..:._~ G,r.ld P. O'Oriscoll, Jr. --~ .tnM.IJf DwII'IO' til ~ W. Micllael Cox Val'tff"""'_ C"""""",Aoto>s ... ... Stephen P. A. Brown ""'""" V",,/'1I$tdrt!l1IIII ~, Economists Narlonal and /i1/ernallr)(JJ1 John V Dvca John K HIli Evan F Koenig RobertT Clall Jcseph H Haslag Kenroelll M Emery DaVId M Gould John H Welch Mar~ A "N'tme KemJ VealS RegIOnal MId Energy Rober t W Gllmet William C Gruben Mlne K Yucel Keith R Ptlilltps lonl Taylor rlona 0 Sigalia O'Ann M Olme(l\ Editors Rhonda Harris Diana W Palmer VllgiOia M Rogers The fconotrnC ftdlr.1 lIueM! Ilowew " pUb! s~ed by 1M 8a~k 01 Dalln [he VIews e>qlIused are If'!ose at !lie aUlhors """ do no1 fle(f$$¥'I'j rellecl !he IlO$OtOOllS of IIle Fe:;Ieril! Reser.e Bani:. 01 0.:1., or tile fedefal ~ "'~ S!mtf ptIOtIS ... M,lable free 01 dl¥ge Plene nn(! 'l!Cllltm fOI .,ng1e-copy and Il'IIItJPe mpy sOOsaoPlOO".s. t.:l. Issues and ldO'ess c/I;JtIgIt$ 10 !he PubIo:: A1la,'s Clr.J¥l. 1"!e!It. f«\PI1IeseM Bn; oIlN'ln. SUtoon K. Oal~. legs 1!I222..121(1651 6289 Art (t$ m3y be repnnIed (Il1I'. fXII'(I,11C1l Iha1 lilt $(It.if(f! IS erediled ....., IN Resulc:h OepaUme<l1 IS prowled worn I toI7f of !he • I Contents Page 1 Misleading Indicators? Using the Composite Leading Indicators to Predict Cyclical Turning Points Evan F. Koenig and Kenneth M. Emery "'111.: U.S. Depart ment of Commcn::1.! composite i ndex of leading indicators (eLI) is a widely ciwd :md intlUt;:ntial economic series [n this article. Ev;m F. Koenig and Kenneth M Emery examine hO\.... well movements in the eLI preclici business-cycle Illming Ix>ints. Using datil that :Icwally would h:l vc been available \0 :1 foreG.steT, Koc.:nig and Emery find that the eLI h:ls provided no reliable advance warning of recessions and expansions Further, in interpreting movements in the CLI, simple rules of thumb have o ften performed as well a." more sophisticalCd forecasting methodologies While Ill!.: evidence in this article indicates th:11 the CLI may provide linle or no :Idvane..: w'lm ing of husint:ss-cyele coming points, the author.; emph:lsile that the CLI maY:>li ll give Ihe e:lrliest avai lable indie.llion of 'l"ch:mgc in the economy's direction. Page 15 A Return to Profitability: The Pelfol'mance of Eleventh District Commercial Banks Kevin J. Yeats In 1990, the comme rcial banking industry in the Eleventh Feder.! 1 Reserve Dist rict posted profits for the first time in five years. Kevi n Yeats exami nes this turnaround and (:oneludes thai hanks relUnled to profitability for three reasons: rhe Feder.!l Deposit Insur.:met: Corporation (FDIC) took over many had loans :IS it resolvl.'(! f:li1ed banks, banks lowered their hurden from nonpcrforming loans by re:llizing losses and removing had lo~ms from the books. and improvement in the regional economy enabled some borrowe rs to ca\(:h up o n del inquent payments. Yeats predicts th:1l if delinquencies continue to decline and net income rem;!in." positive a t Eleventh District banks, the region's ba nking industry can proceed to full fi nancial recovery. Y(.'ats notes that from late 1987 through 1990 the general improvement in the Southwest economy was not rcflcClc..xI in the performance of this region's banks. To reach that conclus ion, YC:It... examines the perfonnance of h:lnks that rc.."Ceived no FDIC a ...;.... istam:c during this time and find ... Ihat Iheir improved capital r.ltio:. ro l1oWl.'(\ the region's economic upswing with a lag of at least three years, Evan F, Koenig Kenneth M, Emery Senior Ecooomist and Policy Advisof Federal Reserve Bank of Dallas Economist Federat Reserve Bank of DaJlas Misleading Indicators? Using the Composite Leading Indicators to Predict Cyclical1Urning Points I.:n:l\ in~ tim •.: I}' nOlit't· of ... lIift~ in till' '::l'onoIllY''s d i n.:ction is impOI't:lIlllO a wid ..., \:l l'il:.'l)' of \:.'<:<m(lillil' ekcisi(lIlln:lko.:l's Ilousc IHlld. . 1ll:IY wish 10 do.:fo..-r the j1un:hast· of . . uch big-tielet'l ilt:ms as :l lIWIHOhik,... :Ind hllll."'\'·.s if l hl.'Y Ix·[it·\,o.: 111:u tllo.: job.s of holt ...o.:hold tHl.'mlx·r.s :tn: in,,,,,,·ulrl.' Firms. .sim il:t rly. ar ...· lInlikd y to \\·anl to il1\'\.·.s1 in ne\\· pbnt :Ind t·quipl11\.·nt if ...al!.: ... :Ire ;lhoUI to dedino.: A pro.... p...·cth·t;: clt:dinl' in sal\:.'s !lli~ht ;11.... 0 inhihit Iht· introdlll1ion of Ill'\\' prodlt<:l.... or dforls to b ro.';11.: into nl'\\' l11:1rl.:\.·I .... For the: policym;ll.:er. bcl.: of lillll'ly information (~111 tran.sfonn \\ h :1I \Y;lS nlo.:anl 10 Ix· ;1 cOUf)[\.·rqTlict! polky into :1 policy (1);11 reinforcl.'s nU{·lu:ui(Jn.... in unl.'lllpl()ymcnt and inlbli()!). A,... lilt' lln ih:'(1 Sl ak' ... sl ip!X'(! in\() :I ro.:Cl's:-.io!l in Iht' ." 'llIlUller of 1990. 1111.: U.S Ikp:lnmo.:l)t of Ult1lm ...·rcl· (.'()lllfXlSil ..., imk:x o f k~ldin~ lIldicil()r.-tht' ).!()\'o.:rnlllL'nl'... prindpal 1001 for prt'di<:ling Im:-.in..: ....s-(·)Th: Ilirnin).! roinl ...--<:olllimrL·d to rt::).!i ... t\.·r ;1 11·timL· hi).!h .... '\101 Ulllil blo.: :-'L·plt:mbt::r. \\'illl th ...· n.:k:I... \:.' of it.... Au).!u.sl \·;tlllt· . d id the COI11]'lO.siIC kadin).! indt'x !x:).!in 10 slitk :--101 u nti l !:ito..: Nm ...·mbe!' did lht' eOlllpositl' index re!(iSlt!f tht' lhn.x l.·onscelllivL· l110nthly del'line:-. th;n are oho.:n l;tl.:\.·n 10 si).!n:tJ :1 ftllllft: slowin).! of re;d 0.:('(1I10111il' :Ini\·it)' Hatht'r Ihall p ro\iding :I(I\';IIKt' \\';Irning of rn·c.s:-.ion. tho.:n. Ihe e()mpo~ilt' imkx of 1000a<lll1g indicllo", f;Iilcd to pro.:dKI ;1 rt;'<:es:-.ion until "ell :ll1l'r Illo..' dl)\\,n!urn in l'(,()llI)mit: ;KlivilY \\ a ... unl lt'r way I i':nllll)lHi ... I .... It) \:Iryin~ dq_:rl.'l' .... ;Ittal'll .s<Im\.' 0 1 till' hlalllt· for 11ll' I"urrl'm Tl'I,,·O': ....... IOll 10 tlw !'>udtlt·n. unO':xlwl·' ...·d I r:lqi im':!'>ion of KlI\\'ail in AU).!II"'!. "ilh i1.... di ... ruptin· imp:ln on world oil markl·t .... :lIld dcprt' ....:-.i n),! l..:fkl·l Oil con ... ulller R UJOfidI'IKl· . Thus. I;'Xl't'pti'JOal circumstances may ha\·o.: 1x·I,,·I) n.~sIX)[)sihll' for Ih..: poor performanco.: of thl.' composil ...· Jt·adin).! in<.k:x in 1990. Certa i nly . a (.';ISllal inspl.'ction of lhe plot of Ihe Il.'adin),! ind ...·x conlainl,,'d in ;Iny re('t;'nt issue of the 5/117'(:1' o/Cllrrell/ HU,\';lIess would Ic:rd one to 1x:1t\.·Vl' Ihalll1o.: i ndex t)'pically <:h:ln),!...,.s direclion wdl in :ld\,anec of pc;lk" :lIld troughs i n Ille ov\,·r.rlll't'onomy. (Seo.: I,\gllre 1. Tht' It·ad lime for pre-ll)l-)() cyclict! P(::I b r.lOgt'S f rOIll IWO month.... 10 \wenl)' month . . and a\'cr:lgcs 9.7 month... At l')'l'lical lrough .... tht:: It::ad liml..' r.tllgcs Ix:t\\\."t.'n onc month ;lnd tcn months and ;t\·er.lgt:s 16 monlh .... ) For:1 numher of rt.:;ISOlb. howcvo.:r. sllch p[OL~ art' mi....!I..:adtn~. Fir:-.t. thc \';lllll;' o f Ih.:: II.:;Kling index for . .'i.:.1)'. Fellfll;UY i... not announced until thl.' end of "larch A onl'-mOnTh :ldv:1n\.'\:: \\':Irning is. therefore, u.s Econ o m ic Mc"ic w - July 1(1) 1 WI? wish /0 /llalik SWpllen P A Brown. Mlchaal P Mermra Ke,!h R PfII/ilpS. Dnd Paula K Tucker for hclplulcommcnrs and dlscuss>Ons We wovld also lIke /0 !hank FrancIS X DIebold lind Glfl()() 0 Rudabusch 101 glVlIlQ us access to mc.r dala $01 MiJllhew I TlJfncr and James L Hedges prOVldtKJ excellent resCfuch lIS$ls/ance Of the four mosl freauently Cltod COIliC/dent IIIdlcalOis of real econotl'lIC BC/lVI/Y. one (r>Onaf}flcultural emplOyment) peakod III June. ooa (constant·dollar persona/Income less rransf/tl payments) peaked III July ooe (COOSlanl(Jo/lar manuilicluring allO trade sales) peaked In AuglJSI tJfld one (Ifle mduS/flSI prodiJCtron mdex) peaked 1ft S6p/ember ACCOfChng /0 Ihs NaIKJNI Bureau of ECI)rl(¥l1IC Research. a privata researCh orQanlZ8110t1 whose flJlmgs on such dates Brit Widely Bccepred the economy as B wflole peaked m July Figure 1 U.S. Commerce Department Composite Index of Leading Indicators I nOe ~ . 1982 _ 100 ". ,, I'T I' T • '" ". ". n. ". '" '" SOU RCE: ~rvey 01 Currenr Busmess. April 1991 ft:'.JlI y no adv:mce w:lrning:1I all. Second. cyclical turning points in the le4lding index are often diffi<.."ult to recognize until well :Ifter the f:1l..1 : Ihe leading index fn:qut·ntly begins to turn up in the middlc of a n."'Ccs.o;;ion--<>r tu m lklwn in the middle of an cxpansion--only to rcver...c COUfSe.: a few mOlllhs lawr. Fin:llly. contributing to the dillkulry in recognizing cydlGl1 !liming points. each v:tlUl' of thc leading index 1.0;; .o; ubfel·t to sLlIlSt:mtial revision after its initial rek:a.'iC. In this artJde. we review the const ru ction of Ihe composite in(h;x o f le<ldi nl-\ indie-.lIors :lnd cvaluat\..· its historiC l1 pt:rforman<:e as a forecaste r of husincss-cyc1e turning points. Scvcral simple methods for forecasting rcce.'isions and exp:tn- Sco, howVvet". DIebold attd Ru<JCbusch (199 1)andHymans (1913) , Of coursc. /tie compoSJ/e 1fIdc)( 01 leadmg ooeillors may. nevertheless. fXOVIdc !he first avadatJle clear !;OI'l'j,ma:lOo 01 a cychcaJ /umIIlg poIn/ , ~Jo icadIfIQ lfld>Cal(¥ scncs 8rea/somalfltamcd by D:l#urnbI<J Urwcrs,Iy's Cont8f lor In/f!ffI8lf01111l 8U$J118$S CycJc Rescare" and. 00"l an expenmontal basis, Oy the NatlOfl8! Butflau 01 ECOflOfTIIC Research (N8ER) l sion.s :lrc considered. We also ('\':lluate the forccastin~ methodology dt'\'elopeu hy Neftci ( 1982). Our :tnalysis dill"el"s from that of other:dlicfly in th:LI . for e:l\.."h mclhod, Wt· nlllstruct forc('a~ts usin~ only Mre:ll-timc" data; that is, we ll.'>t" only inform:ltiOIl that was avail:lhle :It the time thc fo rccast would h:l\'e he..."'{.·n made ~ Our August 1990 foft. .(:;.I:-l, fo r cx:ullplc . i ... ba:.t.-d on hi.storic:d data for Ihe compositt' index o f le:ld ing indicators takt:n fromlhe Al/gUSt 1990 Surt.'I!.V({ C I/IT(!1I1 IJIISilless By COllstn.lCting our fOf<:.·CI:-tS using only infOTln:l1ion tha t wo uld have lx:cn :lv;l ilable at the timc. we hopt' to obt:lin a re:distic assc~ment of the likely futurc pcrfofmancc of the \"'ompositc leading index. \"ie find that rht· lx'h;lvior of tho..: compo."ite index of k:lding indicltors during (he months illllnediatdy prt:cedi ng Ihe 1990 n;:(."t·...sion was not :It att :lIypical. In ~e neral , th en, tht: com posite index ha .~ not changed d i rect ion early enough to pfm'idt' reliable adv:lllCe w :lrnin}: of the o nset of rcccssion.~ and rt.·covcries. \ A reliahle signal of rcct!ssion cannot Iypically IX' cxtracted from thc composite index lUlHI onc month or more after the ro:c\..·:....,ion h;l." alrt':ldy IX'gun Ikliahlc siRnals o f cxpansions ha\·\..· heen availahle with .~01lle\Vh;lt shorter lags In prcdiCling rect!5sions, .~ i1llpl e rul es of thumh havc often performed every bit:l:-' well .I ~ til\..' more .'>ophb(icIIOO Nl'ftci methodology In pn.:uicting expansions. however. thl" Ncftci methodology h:I." had :1 dcar cu~\..· . Descriptio n of the composite leading i n dex All th~ !llcthods for fon..'cisting I.:xpansion.~ and recessions l:X;Jmined in this :lrtide arc bascd on mOVelllt'n(s in the Commcrce Department's compositc index of leadi ng ind icalOr." (CLI). Whi le other compositc Ji..·ading indext!s an.: ava ilahle. the CLI is the only composite leading index cons[meted :md ]1ub1isht.·d by an :Igency of the federal gm'crnlllcnl. ' A:-. :-'Udl. it b tilt." index that is most f:l miliar and lllo.~ t :lcct.--ssiblc to the puhlic Abo. tht' Commerce Depanmcnt"s inuex, first pllhlished in Novl'mher 1968. prcdate.~ othcr a\'a ii:lbk cOIlljX)!iitc indexes. The extemlcd puhlicat ion f\..·,·ord of the gives us :t long interva l ovcr which wc em cV:"IllIah.: the index 's real-time fOf\..'(::lsting performance. eu Fl.-doc r-.. I Resco'c Ib.nk of Dallas fh it:-. 11:1111\: :,uAAI.:.'.' ls. thl' composite index of kad il1).! ind iGltor:. i., :tn :lIn:!I).!am of sevt: rJ.l ul1{krl}'ill).! I.:.'l·onomic l illie sl,: rie.' . e:lCh of which tl.:.'nds to k :td broad 1ll00·emenl :. i n Ihe busines." cydl.:.' ~ T hl.:.' idl.:.':t underlying a compo.,ile index is thaI il will givl.:.' :l m o rl.:.' rdiahk signal of husi nc,....,,cydt' turning po ints Ihan would :lIl y individual indic:llor. Th...· (·omponent., o f Ih ...· CU :m: chosen Oil th<:: hasis of :! <ktailcd ~ori ng sysl ...·111 Ih::11 puts p:ll1lcular I.:.'lllp!lasis on cydic tl liming Other nill.::ri:! im:orpor:It...--<1 into thl' scoring system afl.:.' economic signif'ic:lllCl.:.', (kgrl.:.'e of ('(mfo mli[y [0 the stages of Ihe hu si nl.:.'." '~ (·ydl.:.'. prompt availabil ity of tlK' .~L'fil.:.'s. :tnd [hI.:.' :-.i;w of rt'vhions in Ihl.:.' s(:'ri(:'s. In e:lrly vCfsions of Ihe composile indl.:.'x. the gt"l.:.'all.:.'SI wl'ighls werl' gi\'...·n [0 serit.:s wi th higll ., corL'." on Ill(' basis of IhL' :tI)()vt' cri teria I n Ihe currenl ver.,iOI1 of thl' cOIl1I"x)'''itl' index, 1l00\'l'H:r. L'qua l w eight." are givl'n 10 all the inelL·x ·.' ('olllpo neni series." Tahk I ." hows th ...· :.(: ri(·:, ('urrentl y in the I/.'ading index . Revisions to the composite leading index Tlm..·1.:.' fa ctor1> him!!:r :lItempts to usc the ('omp< ' ... itl.' ind...·x o f Il.':lding indicator,., 10 forec;lst hu,incss-{'yck tmn ing p<)inls: ( I ) tile CLI is puhlbhed w ith :! lag. ( 2) ('ydic !llllming points in the CU :trl.:.' often diffk u lt 10 r...-t·ognize u ntil well :lft...·r the f:l ('\' :Lnd ( 3) Ihe CLI i" .'llbj<."'(.1 to subst:mlial Te vi., ion :!fter i(:. initi:tl reJea~L' In [his scel ion. \\"\:: I'o(.·u.' on thl' third r:tclof, discu .,~ing thc SOllfel.:.' and magnitlldc of J'cdsions 10 Ihe CLI f\l Ihe ('n<l of each month. [he U.S. 1311rt';lll of Economic An:t1ysis ( B[A)-an arm of the COlllllwrCL' Dep:1l1menl- reJcaSl's ;\ value for Ihl' previous l11()l1tll ',~ CLI h:lsl.:.'d on incomplete and preliminary infoL'll1:ltion .- AI the sa me tim ..... Ih ...· 13EA m:!}' :l lso fC" isl' 11K' CI.I V;l lu(·s for :my or all or tile pR'c('d ing fh 'e month., Th..:,e re\"l:.ion.'" (K Cur :t ... ilL'\\' informati on l)C(.·ol11es a\'ailahk on [he U)IlII"X)f1I.:nt seri...·:. A ,' im ilar, hUI mor...· <"Ol11prdK'nsive. revisio n o f till' CU is undenah'n om'e cadI y ...·ar [n add ition. hOWl·\'(·r. tIlL'rl' afe occa:.ion:!1 d la n).:...·~ in the (·ompo ..;ilioll 01 IllI.:.' CLI. as well :!s n,;'\·i:.ioll'. ex {lOSI. 10 tht· \\'(: ights :!II:1Ched to its compo n ...·111 :'I.: n e .. ~ For L·x:lm pil·. in 1989. an in(il'x o f (.'Onsumer eXJX·ct:uio n .. and ;L mea:.urt.' of tk ollomic Rc";cw_)" ty 199 1 manufact urers' orders fOf dlJr:lble goods were added [0 the CU . rcpladng measures of the ('h:mge in busi nes,~ :tnd consumer (,rt:dit and the change in busin...·s,.; im'c l1lo ri es. T he rc v is ion.~ to thc CI.I arc suhstanti:!I. ooh ano..;s pericxb in which d efinit ional dlanges havc (X'curred and within periods with no definitional ch:m gL'S. Fo r instance, o ver [he entire sample period fro m December 196B to febntary 1991 , Ihe ;I\'er.Jge ahso lute siZe of re\'isions [0 [he estimated month-to-mo nth perct!ntage chan~e i n the Cli is 032 For comparison. the average :lhso l u[(~ size of [he initial c.~timat e of Ihe month-Io-month percenl:lge cliange i n [he eu over Ih is peri<xI is 0 43.'1 As w ou ld he expected , for periods in which there hllve lx:t.·n no dcfinitiom!l changes to Ihe CLI . the size of the revisi0l1S is ,~111alll.'r l3etween J:tml:llY 1983 and Jar111ary [9R7. :1 period \vhen Ihere were no (k:finilional ch:mges. the average • BuSIfIflS!l cyCles Ilav91)'P1Caily been defined as sequences of 9XpanslOfl IJI!<i COfltr8C11OfI (ICCCSSlOfl) If) Vi'l/ious CCW'ICI' d8nllf'ldlCalOtS oIaWI1fJ8'e ec;()t1OmIC acllVlty (See llQIe , ) Because f/Ic NBER'S dalUl{1 of buSIneSS cycles IS C'O'I' sldelcd rho most lehllble. lis blJstness<ycle dalesareused If) constlUCI.ng /he C<mmerce Deparlments JeadB1f} 1fldI- calOtS index , FOt a delSlred dcscnp/N){I 01 rhc currenl version of /he COtm'lCrCC Dcpallmenl'S leadiflt) mdlCalrYS IndeX. see /he January 1989 Busoocss CondihOtlS Otgesl & /Ween November 1968 and Febluary /990. lhe CLI w as publ!Shed III me Commclce Dopallmem's Bus.ness COO d,· l,oosOlgest SmceMarch 1990 IfleCLIhasbeen pubiished m lI1eSurvey of Current 6us 'nl)Ss • Important dQf'mfional cnang9s well) ma de III 1969, 1975, 1979 /982, 1983. /987. lind /989 Dcra.led doscr'pllorlsof II1t!SC cllanfjes can be found.n the Seprembel 1969. May 1975 March 1979, Fcbrullry /983. March /987, and Janu' ary /989 IS5UCS ofBus.,ess Cond 'l>OnS OIgest • (Ncr /t)tJ entIre sample perIOd, rhe siandard devUJ/lon of the levrSoon from Ihe ",mal eslrtnarc 01 rhe CLI pctccnlage change 10 lhe fUl81 eSllffl(Jle (as gIVen in Fcbluary 1991) is 45 A~ normal,ly. Ihe /lflPhCalron IS that rf the initial o o cSlmarc IS say a 43-pefcent rncrease If! the CLI (an IfICfcase equal 10 the BVCfage aOsclvle change lI'llhe CLI over me sample pcood}. go percent 01100 IIfTIC the fitlaJ value 01 /TIe CLf will /Ie between --() 31 percenl and I 17 percent Tab le 1 Components of the Composite Index of Leading Indicators Average weekly hours of production or nonsUpeMsory workers, manufactunng Average weekly initial clalms for unemployment Insurance, state programs Manufacturers' new orders in 1982 dollars. consumer goods and materials industries Vendor performance, slower deliverieS dlMuslon index Contracts and orders for plant and equipment In t 982 dollars Index of new private housing units au thoriz ed by local building permi ts Change In manu facturer s' unfilled orders in 1982 dollars, durable goods Industries. smoothed Change in sensitive materials prices, smoothed Index 01 stock prices, 500 common stocks Money supply M2 in 1982 dollars Indell 01 consumer ellpectatlons (University of MIChigan survey) SOURCE : SutWly of CurTflflt Bustness. .lh~ol l!ll· "'1/,: ot dw 1"\.'\ blun ... 10 Illl' l1lomb 10~ month r)l..·rn'111 ;1,l!c cllall),!l' III 1111.: eLl I ... O_IS I" For tilL' "';III1l' p •.:ril It!. t ill' .1\ l·f:lgl.: :Ib~olutl' "'I/l' initl:.1 •.: ...l il11:Hl' of till' f )l.: rl'1.:nl:1).:L' I )1'1:' I/l(' Ixn-lillet/ ·,\/(llislialll'nJ/Jl'''' oJ111t' /Ililia/ (.'I 11~,,'ill/{/I('s ) C I.I , ... (J .U li(~ or lilt: Ru les o f thumb fo r p red icting n.:-cessio n s and expans iuns : h uw w e ll d o they pcl'fo rm? dungl' 111 tilL' rlleJ,I/III;!ry 1987 Bus,ness Con(l I!lOns 0,g e51 nBS used 10 e',II1l.. w tile J,l rlU3ry 1983-February 1986 s,lmple fI,c Feoru,Jry 1986 elate Vias enosen so lniJt tne '''' !'al estlm,l/"5 or me eLI "oold be compared " ,1/1 eSI,m,lieS /I>.ir I>,'we Imdcrgone as many as f!lf!lIe" routme momilly "'vI$fO/IS ""q/', Fa; 11"s /)/'(1(1(1 /he standard deYialion oIlhe rtN,soJS ,1' o 22 A~s.",ul-'9 ,~ltIy the tmpIlCat<CXI 's thaI ,/ Ihelllll<3i CSI"'~"tllS. Sow. " 032-P6Ieern lIlC1e,lS6 Ifl /lie eLI (AA .,lCff"lSe OQI",llu the <lvtll'age /JbsaJute cr.1JrlQC If! tile ~1dc~ CM'f 11111 5,""'1>41 p;:nOd) 90 p;Jfeenl 01 tile I,me /II(' f"~1J eu .... '.e bet-<eon 005 poreelll <lno' 0 69 ~,1I'1O IlIl/IC OL... ct~" nIIL· ... (If "r L'x:lI11i n:l1 i"n ho\\ \\L·II thu mh pfcditt tl lming p,lillt.... ill \\ '1.: llIrn no\\' 10 :111 "''''' \l'1":11 IhL' hlhinL· ...... CydL·. lI~ill~ rl,;t\tillll' (\;ILl 1111 till' U.I TilL' :1...11 .ITlLl).!L' (If thl· ... l· fUll- ... i~ l lll'lr ... illlplit11\ "I hl'~ elll gL'IlL' I~III~ Ix' ll11plt'!llL'ntl'd b~ u~illg d.lL! Ill:' "'Ingle 1"'~lIl' (II till' SIIIH'I'O{CIII'H'III HI/.,il/(~S and rl'ql/ll'L' only ... tt: l;ghtlol'w;lrd 1ll,1111- l' m atilall,lltui;ttil)lb "Ill' r:HIO/l,lk' lor lI~Ul).: fL'a l linll- d:lI.l i... lilat Illl' CI.I , ... dl· ... igI1L'(.1 10 lilfl'Cl"'l hu ... inl'''''-(·ydL' IlIn1111.1.! Ixmlh ThL'rt:iorL'. ,111 L'\;,lu:JIIOIl of th ...· Cl.r~ pI:rli'rln;'llc",' ... l!ould lx- 111.1(k I'm". lh...• PI."t'oIlt..'t11\1.: (11;1 !'or...·l;l ... I ...·r \\ho h.l~ onh Ill<..' tillllll·d inlimn:tliol1 .......'[ ,1\ ,Iil.lhk' .11 tIl<..' lim...• thL' forL'" .hl i... nullL' Statistical Properties of the Initial eLI Estimates Following Phillips (1990), we evaluated whelher the initial release of the composite index of leading indicators has been an unbiased estimate of its final value. That is, we evaluated whether the initial release of the CLI has, on average, been correct. The following ordinary least squares regression was run : Finalr : A + B· Prelim, + er • where Finalr is the final estimate of the percentage change in the CLI from period t -1 to period t, A is an intercept parameter, B is the slope parameter, Prelim,is the initial estimate of the percentage change in the eLi from period t-1 to period t, and e,is the error term. If the initial estimate of the change in the CLI is an unbiased estimate of the final value, then A will equal zero and B will equal 1. The above regression was run forthe entire sample period (December t96B-February t991 ), as well as for the period from January 1983 to January 1987. For the entire sample period, the joint hypothesis that A = 0 and B = 1 could \\ Ink- ,uh,' l'qu"'nl r ..·\ ,... HHl ... tn dala :I!ld dvfini· li(ltl.l 1 ...h.II1.1-:"·' III :I~ ~·IlIl:IIl'''· 111<..' p"'1"fl\rnl:IIK~' ,l 1Ill" eLI ex posl. 11 II 1:--...· r..·\ i ... ,( In, ;l1ld dl:l llg<.: ... :1 rv ,,1 no LI:--<.' in t(u ....cl,ling Il1rnmg point:--. ('.\'(111/.' \\ .... find ll1al. tI,:ll :1 11 :11.."11 in tiK' Ct.I a' III ... igrul lllg r,,·I..",,·,sion ... , ru k ... I\\udl \\"'1.0..:111 1(, lhv 'o il ...• of dnlim" :1'10 (Il<.· n umh"'r of lkdil1\:.' g\'m'l~ril}' do h,,'I!l:r 111:111 ruk' llt;ll I()\'u ... on l hl" Ilumlx ... 01 dvdil1\"'" :riOI1 ..· 1:\<"11 hom llt<.' 1x-.'(-pL'rt()l"fl1illg wI\-... (If Illl1mll, 1[(1\1,,·I .. -r,.1 rdl:IIJIL" ,ign:ll (II fl·\ 'l· ... "on h lIlllikdl 11,1'11.' .11.libhll,' unlil on<.' be rejected.' This result is not very surprising, given the many definitional changes in the CLI since 1968. For the more recent sample period, in which there were no definitional changes to the CLI, the regression results indicated that the joint hypothesis of A = a and B "" 1 could not be rejected at the 95-percent confidence level. These results suggest that, for periods when there are no structural changes to the CLI, the initial release provides an unbiased estimate of the final value of the month-tomonth percentage change in the index. However, whether the initial release of the CLI is or is not an efficient (best available) forecast of the final value is a more complicated matter. For more on this issue, see Diebold and Rudebusch (1991) or Mankiw and Shapiro (1986). $ellal correlalron was delacted In Ihe error lem>s Because an Fleslls IlOt valid ,11M IIfrors we setially oorretaled. the JOinl hypothesis test 01 A • 0 and 8 . 1 was earned out al16f lhe autOCOllela'~n In 1I'Ie 8m)< lerms was ehl!llnated by means 01 the Cochrarre-Orcun meIhod Rules of thumb for signaling recession. r.,in g re l"L"nl 111\)\·..·111\.·111 ... In t il<.' ( :1.1 , \1' ..· <."( m .,idl·" tllrL'L' .'pL·('ili<." 111<.·:1 .... \1 1'1,:' lh.,1 l":11I h..· Ir:IIl,I;lt ...d il1lO I"Il k ' of Illllmh ror .,i,gIl;,li llg r~'l"<..'~,'ions . TIlL' Iir,I- ;,nd ~i lllp k-:--'1 1I1<.'.\'lIr..· i~ 111<..' 1l1l1llhL'r ot" (,>lb<..'L·Ul h ... 11111111111) d ... { !ill "·~ in th ... 'l'hL" eu 'l'l"ond lIIl"a.... uH· i ... IIIL' p1..·rc<.'IlI:lg... difft:r,,'r1CL' 1"<,:IIIL'l'n tlw nlrr.... 111 I:II\I\.· (,I" Ill<.' eLi :\nd its lIla,ill1l1lll \·:lllI ..· (11"'r 11ll" p r~· ....<..'drng t\\"l'ln' '11",1(1)'0 ()IlL' Illl,glli l"IX"Ll nlk·., h :I"L'd 0]1 I l li, 11I(:;I"'lIr..· III pL'rlonn h"'u "'r I h :l n IIIos..- h:I'L'd 011 til ... C IJ 1111111111 ,)1' Illllrl' .,I l,,· r III ..· rl·(\.· ...... irlllll;I' .llr,,·;I(1\ hl"),,:url In ",,}.,:n.lllIl;': ,,·'p.II",II,", no on\.' nil..- or lh~' Il Llt1IIX"r (1IlIlI1''''{'lIlll,,· d L'Lli l1\.·.' in .. I;b'" of I"llk·:--. 1);)':1 d ..".lr :ldl.11I1:Ig..· 01 ..' " oth"'r, hL·.. :tll'<.' \\"'ighl " pl.,n·d 011 111<..' siz('of (k-d ill\".'. not IlI'l tlldr illlllllx.·1' T ill' Irnal 1I,...;hur..· is til<.' I !t·r..· .1,g;lill . rL"lh ... liLIII~ . III ..· CI.I unnol Ix: ...., . 1"1\.'( I ..·d 10 pn'l id ... n.-li;ri,k .Idl ;lno.: \\',Irnin;.: ~'Ild of :l r"'cL""illll f. ... >1I<>mk ReI';"" - Jul)" 19'1 1 "r lit..· p ... rcl'ntag1..· g .lp 1x.·nll·l·1I IIIL' nlrrl'nl \ ':III1L' of 11lL' eLI ;,nd a nl..-lll· l\Iol1lh llIo\'il1l-: alcr.lg..· of p:,,1 1·:tlW..·., FolIll\\ ill).\ \11 M '1'... and Z:lrnO\1 il l: ( 11)1-11). , Figure 2 Predicting Recessions: Performance of Consecutive-Decrease Ru les Number oIlalse SIgnals 01.11 0121. monlhs Olilpansion " " " " , • ---- F.w_daII I Average lag {months) Wl' nmve.:nt!d this gap to an almualizl'd percentage.: r.lIe of gro\\,th,l ! Each of tllt!sc thrtt rnC;I"lln.:S Gin he comhirn:d with :.t cri tic:.tl \'alllt! to ridd :1 nlit: of Ihumb for pn:dit:ting the onst:1 of 1'I.'l'{;s... ions, For cx~mpit:, $pec11ic811y we diVIded the gap Oy 6 5 10 converl II 10 a monlhly prOWlh falo (the aWllage of/he plccedong lwelve monlhs IS iocatoo6 5 monlhS before thecuffenl monlh)and lhen mullopl,oo by 12 I.' A signal was mlerpretf1lias falsoo! 1/ was rf)VCISoo before Ihu onsct 018 rocossion For c xamplc, usong tWO consecutive declines In tne CLias /I Signal of fBCesslOfl, we counled a signal as false " IWO COflsecutlve declines wero followed by a(l IflCfC/lSO beloro e rccession doveloped Once the economy entefed a recesslOfI, as defined by the NBER, tne<8 we<e no lalse signals , Becauso for Instance the Avgvsr CLI value IS nOIre/cascd UN" the end 01 5eprember I srgnal 01 fCCCSSIOll IhallS based on rhe August CLI W(ItJ/d lag one month for a fcccssron t:IegtntwIfJ "" AvgvSI For rile final da/a. we hal"8 CO'l$CCurrve f'I(]f pIoItfK11 potfI/ for four mon/hIy dec/Jfles in Frgurc 2 becausc four CO'l$CCullVtl' declines did IlOl occur before or dufng the 1980rccCS$lOl'l The all'8fagtt lead fmc IS, therefore, """rotc n IflrS casc 6 using tht' (:onSL'('lllivl'-(ll'lTe;L..e mC'I."llrc, on\.' mighl inIl'rprt't two <:(m~L'l'lltivt' dt,:din\.'s :1:> :1 sign:ll of n,x~'Ssion Allemati\'dy, one might require thrL'I.-'or four--conl'(,;cllti\'l' ,!L'din\.'S a~ a ::.ign:ll of rt.'(.'t'..sion In gcnc1.l1. thl' lHOrt.: sirin/ott'nt tht' crilic:11 vallil', the Ie!'.... ;I(h~ltl(\: w:lming of ll."Cession :t gi\'l'n llll·:I:>Urt.· will prn\'idt' At thL' same time, the.: moll.' ~rin/ott!nt the l 'rilil':ll \':,Iul', Ihc fe\\cr faL-;c ~ignab of 1l."'CC..."joll till' 1llt':IMIIX,' is likely 10 gi\'t' Fi/oture 2 pll.'SCnts thl' po:-.t-l968 perfonnano: of nLlc~ ha,'o(."(! on til\.' numhL'r of consecutive dt!dines in thc CLI. Thc V\.'lli(.':ll axi~ shows thc number of Illonths during whil.:il eal'h rt.lle.: gavc;. f:l lse signaL " Till' hori zont:11:Ixis ,~ ho\\'s thL' ave.:rage lag (again, Illl'asured in monlh,,,) with which t!:tch rule sign:lled lhe fin: 1'I...'(.·\.'s,.. iol1s in lhe.: sample pt:riod,' The pcrfonnance ()f the C()11'''e.:c.:Il1ive-dedine I1ll':I."lIre is shown bOlh for the n';:II -limc d:l1a (data I h:lt wou ld havl' 1)Cl'n ;I\':,ibhle 10 :1 forcc:t:-.tcr) :Ind fo r thl' final rt'vi,..;cd (!;tw, For in::.lance, with rt':tl-tilllt' dilla, if two con,..;cculive dl'('ft!a,"L':-' in the indt'x of It!ading indiGltors had !xcn tah'n :lS :I signal Ihat a r\.'(·\.:~si()n w:.ts under way,:1 \'alid ~ignal \\'ould h:I\'l' ('ome, on :I\'Cr;I/otL', 11m_"'\.' monlhs rif/<'l"lhL' n..:n'..sion had ;Klll:llIr hc,:J,:un [n (;iglltecn mon lhs (oUI of:.t lotal of 21 ·1 ll1onth~ thc economy W;lS l'xpanding), thc nllt' would h:I\'c i.<isl!(.·d a Ltlsc :-.ignal of n."'CC.....ion If thrt'l', r.llhcr Ih:m two, ("on'oC("Ulh't' dc("n.:a'>(.'s in lile i ndex had 1>Ct.'n l;tk\.'n :I., :t signal of rl.-'(.'c:-',.. ;on, ;1 \':Ilid Signal wou ld han' arri\',xl j H l110nths I:llt', on aver.tge, :tnd fal...;c Signals would have hL't!n gi\'(::n on fiv(' <KClsion:-. If four con~\.'('lI'i\'l' incre.:ascs h;ld 1)Ct'n rl'quired, the sign:11 wOlild havt! 1)Ct!n fivc months !:tte and llevt,'r f:l bc _ A~ \\'ould he t!xpcclCd, the tfad e-off between fal."e ~ign a ls and lag limc appl'~jrs 10 he more bvorahle w hen final revised d:I\:1 on I he eLI arc used r.lthe.:r lhan lhe data ava ibbll' ;II Ihl' time IhL' f(lre.:cl~1 was relevant.' ~ \'\11..' next \.'xalllinL,d signaling l1I1cs hascd on the pen':l'nlagc ditTcrt'n('l' ht'I\\'t'l'n the CliTrent \':l llIC of tht' CU ;Ind th\.' maximum \':tlue of thl' eLI O\'cr the pn.'n'<.ling t\\'cl\'c mo nths, For such rulcs, w .... found th:tt tlll'rc would >;Qmclirnes h:I\'t' 1)Cen ;In ;1(1\':lnl:lgt' to ignorin~ thc mosl recently rdea:.cd \'alue of tht' CI.I That is, there was sOl11etimes so much noi"l' in tht! initial CLI dala that it would have Ix:cn IlI..'tlcr to wail one month, for the d:lt:1 to go through a n'vision, ralher than FroeMlI Resen 'c Hank o f Dallas Figure 4 Predicting Recessions: Performance of Decrease·Relative·to-Moving-Average Aules Figure 3 Predicting Recessions : Performance of Decrease-Aelative-to-Max imum Rules . Number oI lalse !lIgnats OUI 01 21 4 montns of expal1$lon Number 01 false SIgnals out 01214 tnOI'lttts 01 exparlSlOtl .. -", - ~_lWIta :m \_1' ••.•. -:. \\\\ ~ (In! re.<I$IOIl) ~_ • ~_('Int-' • " .' \ -?'110 •............... ... .. -, -----... """ ..... <o -I"II. • .• •• ) Re.oI.~ ~ ······F.... .. • ." o \ 2 3 A , . . lag (tnOI'llI'llJ ... • , , • I~LJlgl" .Ipply Ih~ 'Igll.dill).! rull- to till: d,11.1 il1l t i.lll~ !ltllllh... r <II I.d .....: ,ign.tI, rd~;I,\':d.' {;nll\,;11 1,lIltl.·' 411' -I pt.·r,'\"111 through (1I IIh.1 1'" 1II0nih .11\.'l":lg ...· k·:ld l i l11l") to I... n (wi[h - ! pl.'fl.'en! II \,'rl.· l.·otl"dI.Tnl I I).!Ufl.·.~ plol' IIw .\u"Ordlll).! to III\'" Il).!un·. II 41111.' h;ld u ....:d :1 rult.· h.I'I.'<.1 on Ihl.· Iip"l 1'\,'11"011 011111.'" <:1.1 :md .1 l..,Ii(;11 \;lllh.· (II - I 1><''I"I.I.·nl. I;I I,\" rn\""llln ,iglul, from [WI.'IUY-[\\"() .1 I !.·lllllnlh .11 ...·"lgl.· 1,lg) 10 onl.' " "l h:'I ·1 H·11l0nth .111.·I<lgl.·l:Ig) r...,ult;1111 IX'I'i()fIll,li)l.l' 1"1111111' , \\ \I.·f~ 10\\ IlUmhl.·l..... 01 Ltl,t· 'lgl1.11-.. '01111.'\\ h,lt 1"'.:UI.·r \I h.ll ...· hl.·\,·11 .It:h,i.:\l·d h~ :.llItng 10 'I.'c;' tl\t:' fir:-.l \I ;Irning l illl<;'... could 1"l'l I'lon 111 Ilw CU d:II.1 \\ ()uld h;11 I.' 1""'1.'11 11111.lil,,'d til 1\\ 1.'111\ -1\\ I) 111III1Ih,. ;ltld.1 1.11,1.1 '1).!lul \\ould h:l\\.' (t11ll1.·.!l1l .II\,·CI,\-:\,·. 06 monlh 1""'j(lI"l' ;1 I"l·ll.·"'OI1 .KIII:dl~ 1X').!.11l \\ IIIl .1 lnt lt,1 1 1.lh ...: of 2 jX'rcl.·lll. Ih ... rl· \\ou ld 11:1\'" h\.·\,'11 '1.'11.:11 fal,\.· ,igll:d, .. 11lt.! .1 I ;lIid "glul 1~ ptc l l1~· II ()ult! not 11;11 l' h ...·\,·n oilt:lilll'd unl11 I !. t11()nlh~ (~ji(,".1 1"l·I."I.,,,,OIl Iud .lcl\l:lIl~ hq.,:ull ()Ill,: had \1;II1I.·d f(l1' till' fir'l (:1.1 If rl'li,j(JTlIII lkdill\.' I IXTl l.'lli r<.:l.ltlll.· to 11K' C I.IIII:I." jIlHIlII mnllll· pn':l."l.·ding 1\\ l·11\.' 111 () 11l11 ~. ( )lll' 11()uld It;ll\.' oh.' I.·rn:d no f:II,,\.· ,ign:tI, lIuI \\ou ld 11111 !t;'H' found ou t :Ihout .1 r\.·C\.·"1011 1I111i l "'i '" llHllll it.,. on :1\'I.'fag ...•. :1111:1' it had hq.':l1l1 I .\g.11I1. 11t~· fillal r ...·1 I.'t·([ d:lt:L 1>:lil11 .1 nll~lt.-:l dingl) f.1\ l1r:IIJk pktll1't' of th\.· Il":tt!...· off 1""· \II I.·I.·n f.11."i(.· :-.ign:tI, ;lnd l:tg lillll.· Figu1"i.: ! illu"tl":l!L·.' tht, r><.·rtilnll.uKI.· of nlk, \);1'1.·(1 on thl·/-tap 1""'\\1""1.'11 I hI.· nerr... nl 1:,lu ... of lilt.' CI.I and .1 I 1\ l.·hI.'-Illonl h 1110\ tIlg .1\ ... CIgl.· of p:l:-.I \a llll.·' A" \\1111 Ihl.· PR'I i(lll' (111) fi).tur... ,. I.·;tch poinl in Iltl.' dbgl'.IHl 1"l·I)1"I.·'I.·n", Ih... pl'l1or- 1/1« d,aor.lf1t l<1J<es mto account me add.llOflai one·mOll1n ~10 ,n~ Nl ,..,·,MQ lor /flellrSI ":'!VIS/Ofl oflhe eLi FOt o/hr!f Slgnal,ng rules consldo<ed In this article. "';lII,ng IOf reVl$IOfIS to IhtJ CLI was generally 110/ WOtrh ...~·"le (111.11 IS the loss of ()tJ{1 mOtllh of Icad t"lle more than offsel .my rfKIUC/1()tl111 /he lIumuer ()lla/sa Signals) FOt rIi{l "nal valvas ()I me CLI. Ille average lead t,me 1$ 4 percent For Ifle r~:J1l1lt1l) dill;} perioilTUlnc8 POIntS c()uespond· 1I1"III!y l()f IhtJ cfIIlC:J1 value 01 "'910 tile cflt,cal vallJCs 01 2 percem and 3 percenl are flOt sllOWI'I bGcause they he 800ve a Imeal cCJmo.natlon oJ me pom!s corresponding rolflecnlic8/values 01- r percen/ and .. jJOfccnl A /or8ctJsle< could nave done beller by choos· Ill{} IIVS linear combrn.1lJOfllhan by choosing e,rher ·2 per Celli Ot 3 percenlas /he cfl!JCal valt.e for hIS $lQfJahng rule In clfCCI. Ille "'Je$lOCPcale /he Oesllrade-<Jlf /his measure could ha~e )I1CIried Irt prt'd/cWIQ II!e past f'V'fJ recessl()tJ$ Ag.l1fl pOtfllS cor'esponOlIlfJ 10 some oIlt1e Cfl/Ji;a/ valves m,tntl.· of .1 ,"gn;t1illg (uk Ih.11 ('1Jn",'p<lnd, to :1 are 110/ ~It'd becal.lS8 these poIflIs be aoove a linear d ifTl.·rl.·nl niti...:.,1 1.lllK' For "':'Il-tililt.' 1.1;1\,1. lilt.· COtlonallO'lolOfher feaSible tJOlIlS and so. f...:tl " Olll ic Ik"lt-w - Jul y I 'J<) I afe~led Figure 5 Predicting Recessions : Real-Time Data Number 01 false signals 0111 O. 2'4 monlhs 01 expallSlOll Col ...... _ _ f\oIIoI ~.. . ·····Oio_ ...... -.-f\oIIoI(.... . - o ) -Oio_ ..... ..... ~ " , ............ , , , 3 AvEtragEt lag (mOnths) '" • , , , l'1~lln;' .:; !'>Ij{)\\'" Ihl..· n )lllp;'r~I{ , \ I.: p ••:rl( mll:"ll'l' 0 1 .hl..· Ilm:t.: d i ffl..T~'nI d:,.........,:-. o f I'\,k·!'> of .humh u:-.in~ n.:al-liI111..· d:II:, frolll Iht.: p:'",. fhe l"\..'l·I..·:o..... ior,.... T he: HlII..-:-. 1110:-.. oflt:11 1l]l'1l110,lo,:d in the popllb r prl..·s,-tIhn......· h:':o.(.·d on .hl..· 1ll1l1lIll..'J' of nln:o.o..TlIthl..· d ..:dinc!'> in Ihl' ,mkx of 1t':.din.1.! indic.llor-.--:,I"\..· no( I} pic..I1~ .Il\..' ,op rX:rfonlll'''' t ·nk· ...... Olll: ;:-. \\ iIlltlg 10 tolt',:llt' :. Ih "'-month lag be1\\ ,,'l'll !I,l..' linll..· ;, rn'c"",;oll ho.:gins :md tht' .illle :1{ which it i... :-.;gnak'(I, Figul"\..' .:; :-,u~"'''h .lIal orll" (:ould gcnl·f:.ll1y <.kJ 1-1..:111..'" hy lIsing :1 fUll..- ha ......'(1 Oil till.' dilh:, ....·nl'l:! 1>,,'t\\·I..,,'11 ,I1\..' CU ;md i .... t\\dH:-l11olll11 IlUXll11l1l11 or h~· 1I... ing a n.lle Il;I.......'(1 011 tht: ~ap ht.:'\\I..·~..'11 Iht' CLI :Uld ih 1\\1..·1\'1:ll1on,h l11o\'illg ;1\ ·I..· ra~l..· ·1111..·-.c;: 1\\'0 l'h.........·... of J'ult::-. h;l\I..·IIK· ad\':Ull,lgl' of payinJ.l;l'" Illl lch :1II1..·ntiol1 10 th .... siz(' (If (il:c hnl..·:-' in 11K' k'ading index ;IS to Iht: /llIIl/lx'l'o!' dcdinc .... hu) for l11(· ... e dassl..':-'. ho\\,e\·"",. on ..' ..':111110, ••:xp\.·{·t to reed\'l' any :,d\':1I11'e: w:llning or rl..'Cl's:-.ion ... 1lllks:-. onl' i.'i \\ il ling (0 In prcd/ClIIIf1 c~pansJOnS_ /he mmrnum and fOOVJII9.aver age ruleS PClformed bellf}( \\'hen /he current eLI value was compared "" rh SJII paSt ilalut's than ,,'hen tile CUlrenl eLi value W<JS Campilfed .... ,1Il ",oeive paSt iI<J/uOS A posSIble eJlf)/anatlOt'l IS thaI rile average Jengrh aI me past lour recesSO'lS IS roy 122monms In iorecaslrng recovf!flCS unhke IotIXiJSllfIg ,ece$SlOr1S "'iJ1rrng /I mon/fl 1011/16 eLi /0 undergol/S f"Sl rcvr:>iot> does fIOI • entIiJIICe me periOlfnance 01 any oIlhe rules ,ok'r:II ..· a ... lIh... t,lIl1ial 1l1ll1l1x 'r or 1:11.....· ... ignal ... \",fhi k . hi ... '>(Jllnd.... 1,1..\".1 I)(IIJI' pl·rhH111.'Ilt·..·_ ... tldl mk-.... IlU~· . nt:\\.·n lll'!t.-........ pfl.l\kk thl..· earl' ..·...1 ,I\'aibhlc \"onllnn:llloll tlul ;, r..'l·\.· ....... jOJl ,... lind..·.. \\;1)' Rules of thumb fo r Sign a ling expansion . I(uk·... of thumh :lIIaloj.l(lu", to tho ......· u ......·d 10 rOfl,;'G,!'>1 rl.· .. ·\.·~ ... ion!'> Gill Ix' l""t:d to ""l-!Il,,1 f ...·CO\ I.'nt'~ For in ... tarKI..·. on\.' LIn u .....• Ulll ......·cutl\ .... ill( fl·;I......." in Ihe Cl.llo ... igll:,llh..· h.... ginnin~ of an ..·,pan ... ioll /\h,,·rna,h·dy. 011 ..' can look al tIll' difli:n..' IKI..' h ....,\\ el..·n Illl' n'n .... nt \ :llul..' 01 Ill\' CI.I :Ind it.... minimu m \':du ..· 0\ "'r ~OIlW n.:n·,ll 1I11 ..T\·al o r tilt: gap l1el\\ .....·11111...· uJlJ'\.·nl \;1111 ...· 01 . 1.... eLI :,n<.l a '(11)\ 111).: :1\l"~I).:l· III p;l...., \;tlll .., ... til 1)111:(in:1 :-.;gn:11 I 11':111 UPU)lllll1.1-t l·'p:ln:-.ion 1'1 FiguJ'.... " (I lhrou).,:h 9 :-.11(1\\ th ..· p .... rforlllanc.... Figu re 6 Predicting Expansions: Performance of Consecutive- Increase Rules Nllmbef oIlalse SIgnals QUI ot 55 months 01 contractton , • , , • • • of IIl1.' ,hrt:l- typ\.· ... 0 1 rulc:-. o f .hul11h for v:l rioll.~ n itic:ti \·;l lu ...·... O\\"r t h ...· fOllr rl'~'()\ ... l'il· ......incl' 196H 1\ ... \\:1'" thl:.· C.ll'o\'· in pr\,·dil·t inj.t .....·~·l' ...... ion... u~inj.t fin:ti rl..·\ il'ot.·d d:II;( ).:,\'1,.... ;I \·el1· misll::llilllg piclllrt.: of Ih ..- p(.'rfonll;IJlI..'I.· ollh\:: mil..-... in ,....:d lim\.:" :O' I~ull'~ h:Is<.'d on 11ll' illlllllx:r of nm ......·nlli\·\:: inn...·:! .....·!'> in th ...· CU or on Ihe inH\.·;I ....... in Ihe e LI r ... l:1li\·I..· 10;1 J'\.'(·\,·nl IlHnimUIll 11:1\\,':' ... Iiglll pt:.'rform:IIK ... ",'<1,::(1.' 0\\.'1' ruk· ... h;l"I..- d on Ill\.· g:lp IX:I\\"I..'C1l Ih..· nnTt:n, \ allll' 01 Ihe C I.I :lIld I"" m(ldn~ Figure 7 Figure 9 Predicting Expansions : Performance of Increase-Relative-to- Minimum Rules Predicting Expansions : Real-Time Data Number 01 false signals out 01 55 months 01 contraction Number 01 lalse SIgnals O<J! of 55 months 01 contracllon , '" '", Real·time data ...... FIftaI re .. seo data •\ ,• ,• , ......... '" -...... ,• , , .." ~ ..... ~ o§ , 1.5:2 Allerage lag (months! Figure 8 Number of false sognals OU! 01 55 monlhs of contractIon • , ,• • .." - , 05 Fleal ·~me data F....I f9Yised data 15 Average lag (months) Consecuttve-increase MIs 1.....ease-<.Iiv-e-ICHllir......... ruIe$ --_.'_~ru18s , , , , Average lag (momhs! Predicting Expansions: Performance of Increase-Relative-to-Moving-Average Rules , , -, - 2 " :l\"er:lge. Still. nom' of the f uk...; CIIl he relied on to prov ide :ldvancL' warninl-\ of exrxln,~ions, Arc sophisticated signaling rules any better? The Neftci methodology , , • hul. in rrincip lL-. m:lke," IX:Her USl' o f tile infonnation a\'ail:thk· to the an:tlyst.!1 As :1 practical lIlatll'r, \\ 'l ' fin d th:ll \\'hih: thl' N'dici methodology is c learly superior to O llt" ru les of thumh in pn•.-'dkting 1.:xp:m"ions. dll.: rl· is nothing 10 Ix' g:lined by using lhe :\'dki appro:ldl in predicting recessions, ThL" ~dki melil(Klology assUllles thattl1(.· hislOry of lhe eLl can Ix' divici1.:.'{1 inlo exr~lnsi()n ~Ind ('( mtl:Ktion pha.s\.'s, \\ ith lht.· hehavior of the index d uring i t.~ 1..:xpansion p hase lX'ing fundament:ll ly d itl"cn:nt frum tlw IX'h:II'ior of the index durin~ ils uJIllmCli(lI1 phasl', Till..' Ilh.:th(xl(llogy furth er :t"Suml',"; tllat [liming IXlint,.. in the eLl \"ill pfl'Cl'de lurnin~ ["Hlints in till' ()\'('I:tll li,."onomy, T he ho r)(: is 111:11. u,, ,ing a fortll ub <iL'I'c1opt'<.1 by NdlCi, thl' :H1;tly,~t will Ix: ahk to recognizt' tha! the CLl has entefl'd ilS umtrac(i()I1 (e''''pansi()I1) pha,"i1.: lx-fore lhe l'conom}' as a wholl' lX'gins to ded ine kxpand). 11' the Na tional Bureau of [("(lIlomit' Ik:->l'a rch ( N BEIO dedares Ihal ~, recl·s...;ion endl'U :IS of ;1 ("LTta in month. till..' Neftei mcthodology scts till' proh:lhil ity of n.'ce,~sion eq ual to Zt'fO in that l1Ionth , In SUhSL'qul'n t months, the Neftci formula is 1I...cd to CV:tlU:ltC l'vidence on whether t he CLl is in it...; contraction phase , The evid enct:-W tl hering Neftei (J9H2 ) h:t" del'c1oped:1 llletil(xi for J()f(:cl,~ting husinL'~~-("y("k' turning ["Hlints tll;ll is not :tS e:l.~y 10 il11pk:l11l'lll :tS o ll r ru lL-s of l humh f.conomic Re\'iew - Jul y 199 1 For an mternatlonat apptlcBlIOfl of the NeflCi methodology, see NiemrrB (1991) 9 Figure 10 Predicting Recessions and Expansions: Neftci Probabilities . .." .." .." p"""" , " " "o p , --p ., p '"- , P ..... .~ I1II1 W~ pnX'l.·...s is nmllll:Hin.·: th..:- prohahility th at the ClJ i:-. III it ... contrac tion p h ; l~e in thl.· current month (;md. hencl.·. that the l."(·onomy i ... ;Iholl! to enter a rco..... "ion) i .. an Hl cre:l si n~ funnion of the proh:,hilily Ih:u Ihe CI.l \\":1:-. in il:-. n mlr;Kl ion pha.,e d urin~ th .... previous lHonth Th .... proh ahility Ihal the CLI i .. in it .. cont r.lctlon pha"l: in the \.."ll rrent month i:-. al ~o a funnion o( thi .. m on th·:-. I: han~e in the imk-x: lhl.' 1:IIW:r tIn- d en e;lSl.' hm:l ller the incn... a.~e) in Ihl.' CU. Ihe greater hi l he prohahi l ity tllal the ind ...·x is in il." (·o!llr:u:tion phase Fin:lll y, thl: proh:lhilit y tit:!! the CLI b i n it .. (·ontr;lC"lion phase in thl: (·\lHt'nt mOllth b an innt':lsil1g fUllclion o f the numher of IXlst (·ontl~t("ti()n ph:lst's n... hlth·l: to the Cll!llLlbliv ...· [l:ngth of past t'xpansion phaM..·s: if transitions frol11 ...·.'\pan .. ion to com racl iOIl h:I\'t' hee n ran.' in 11 K' IXI..,t, Iht' Neftci IO]"1nub aSslllllc:-. th:!1 Ihl'y prohahly w ill :11...0 hc r.lrt' in tht' futUR·. (Fo/" d eft/ils. seC' Ibe box 1i11C'll ""he X(dki POl"lllllla .. ) Once Iht' \, BER h:l., (k-d:m:.·d Ihal the cn>nolll)' pt.:aked i n :I l·crt"in l1Iontll, a procl"tlurt' analogolb 10 tbal olltltnl,.xl aho\'e (·;tn be lISl.'d to t':1!c1l1:1Il.· the prohahility Ih:1I Iltt' CI.J is in it.' cxp,tn~ion phaSl' (:Ind. llt'nl1..'. th:1I tlK' ",'COIlO1ll)" i.~ aholll to r ...'l'o\·.... d Figur(' 10 plo" Ihc prohability of imilli/lelll rL"(.·cs.~ ion ((Iminl-l pc.:ritxl .. \\"IK'n Ihl' economy was 10 IN " .,. .,. .. fl W . . .... cxp:mding) :md thc plllh:lbilit), of imminent t"xpansion (durin)..: pcri(xb when 11K' I,.·conomy w:tS COIl 1r.lLti n~ I ohtained hy :lpplyin~ tht· :"Idtd fomlli b 10 rl:a] -time data TIlL· tlftlll\: ,how:-. thaI Ihe p rohahilil)' (If immin.... nl rt.·(."e:-... iol1 \\':1:-. :1' or abov..:- 90 pt!r\.·l'l1l-1h.... nilil"al value :ld\'{x~lll..d h~' Ndit:i-al tht' ..1:u1 of only Ihn.--e of Ihl' IX1..1 fi ve rlxe:-....ion .. FlIrtIK'r, in 1\\ d\'e month." Ihc prd):lhil ity of fl"<."t:Ssio l1 n::1l'IK"t1 :11 1e:I..t 90 pt't'1....... nt. o nly to bll helm"\' I)(J peR·L·nI \\ithOllt a I"ccc:-....ion. Th:1t is. tht' Ncftd (Ol11ll1la gal e 1\\ l'In.' (:11:-. .... :>Ignals of fl"Ccssion Similarly. sull lI:-.illl-l a 90-pcrn'lll crili cal \·:r lm:. thL' Nc!ici forlllula g:lV\.' :tch-ancc or coillci(il-nt \\"arnin~ o f onl y tWO of lilt' p;I.~ 1 four n::cov(:.'ri ...·s In one monlll. IhL' formula ~i.gn:lkd :l rt'cm·cl)' hut rl'vl'!"scd ibl'lf Ixoforc lh ...· recovery hcg:Ul . Fij.\url,.· II ~hm\'S thc :IVl:r.I).:t' I....:ld lime :Ind n umhcr of f:t l ~e re(·l's~i on si/-lllal... .gil·":-Il hy thc \'dit:i fonnub. u.,i ng \\\"o ,t he fnall\'l.· c ritical \·:II11e:--70 pe!"ct'1l1 :lIld 90 percenl Tradl'-ofb hc.:lweL·n 1t':ld l illie.· :Ind f:Jbt" .. i~n:d .~ are plotted. l .... n~ hoth fl.·:II -tillll' :md fi nal dal:1 A~:lin. our l'()nlt~nllon i..; ,h;1I the 1r.ldt..... IlIT oh{:l inl,."'{1 hy u:-.ing fi nal data i:-. miskading 11)(:.' tr.l(k· . .off \\"1,.' arc mo::.t 111\1.:'1 )" 10 f:l("l,.· in prl'di L'li n~ fu l ure fl.'ccs .. ions is I hat ohtaincd frolll rt':l l ·l lIlll" dal a. Fi~lIfl' 12. \\ hich (·omhines FigllR' II wilh Fi~un: 'i. shows th.1I u .. ing thc Ndil"i rlll"lhodology F.,m-ral R ~n'c: Hank of Dallas FIgure 11 Figure 13 Predicting Recessions: Nellci Formula Pred icting Expansions: Neftci Formula Number 01 lalse Sl{jl'lals OUI ot 21. monlhs ot elpansoon Number 01 talse sognals out 01 55 months ot COl1lractJon . .- --• ...... flnII_ - , , ANl-___ -. "" • ....... ..............I • •••••• ArlllOIII; ~ , , --.3 ~2 -I 0 Average lag (months) 06 -<14 02 ....... ....... ..................... .. ..... &0'11. 0 02 04 06 1.2 08 I. Average lag (monl hs) Figure 12 Figu re 14 Predicting Recessions: Neftci Form ula and Ihe Rules of Thumb Predicting Expansions: Neltci Formula and the Rules of Thumb Number 01 latse SIOnals OUI 01 214 monlhs ot elpanSIOl'1 Number 01 'alse Signals out ot 55 monms ot COOUaclion ,,-~- ~~•. ••••. o.:r-·~·IO·_II.IIe${~ - o..::.--_-ID--mCMng __ • - NeIlD Ior.....,g, ,eal-llme diIIao rev} -- ~-n:.-...........1r;>--rIWWOUn1l.lle$ - n:.-. ...... ~-.n.tes n.tes -- ~IiOII"I1oJ6a.""""'_ , " s ..... ...;] -2 a. .......... . -10123 Average lag (months) • • , , 10 pn... diu rr..·('I.· ...:.lon ... r idd ... Ill) !-::111l i n p "'rf, .r· m.IIK ..· 1"I:.. b ll\·..• 10 u ... in!-: ... impl.... ruk:. of I hu mh. Til .... 1"lI.."1"lo1"m:ul ("l" of '11 .... ' did m ....' h(ld()lo~) III lor....G ... ,inl-! r..·"p;ul ... ,on.. i ...... ho\\ n in Figlln.: I j 1\;"::lil1. Ih..· .lbl l ,IY of tl\l." lII.... , hodo l og~ 10 fOR·l·:N h tl .. il1 ..·..... (}l·IL·llIrmnj.l ])o;n,... I... l·'\;I).:..l-:l·r:lI,,·d \\11"'11 1111.11 1"l"\I"o(.·d cbu :lrl" ll .....·d I .. mg rl":1I- l ill1l.· <1:11:1. 11K' "...-titi nlr..·lhodol()g~ C lIl IX" hl'lIn dt: ..... nh(:<1 :1 ... :. l:oincid"'nl ind'Gllol" oll"\p:m'lCm ... !lu n :1... :1 1..':.<li11).: i ndic:I\()r o f l'''P:II1 .. ;o n ... F..co ll o m k Rl·V;"-", - Jill ), 199 1 -, , , o , Averagllag (mooths) • • From Fi ~ lI rL' J I. ",hidl com bi lK'" Fig lll""-" 15 \\ 1111 I'i!-:un..· 9. \\.t: ("~11l Set: tll:LI in pnxlict;n~ ..·"p:! ll . . iIJll .... thl· Kdlr; flJrlllllb ()Ullx.. rf(lnll~ any (If 11ll" rul..-... ()r Ihum l) t"()lhidcl"cd :Ilx)\ "'. SUmnl<lry and conclus ion !'I1.lIlti.lft! 1..-••dlll).: plot .. of Ihl" COm~J l e IIltie" of lIld lCilOl' ~i\ ..·:1 ;1I110UIlI 0 1 lll"lp till" eLi lIu:.lcading picl u re of Illl' em p1"O\ ide: the "'("o noillic . The Neftci Formula The Nettci formu la for the probability of recession is PI:s (PH + PL (1 - PH) ] F~ I{[PI _ I + PL (1 - PH )) F~ + (1 - PI-1)( 1 - PL ) F~}, where P, is the estimated probability that the composite index of leading indicators is in its contraction phase at time t, so a recession is imminent; PL is the a priori probability that the ell has entered its contraction phase, given that a month earlier the e Li was in its expansion phase ; and F~ and F~ are the likelihoods that the latest observation of change in the e Li came from the contraction phase of the index and the expansion phase of the index , respectively. Following Diebold and Rudebusch (1989), we made three adjustments to the original Neftci paper. First , we simplified the Nettci formula by using the Diebold and Rudebusch (1990) result that the a priori transition probability (PL ) apparently is not, in the real world, a function of the current length of the expansion . (In practice , then , Pl is set equal to the number of months in the past in which the economy made the transition from expansion to recession, divided by the total number of months in which the economy was expanding.) Second , we assumed that changes in the eLi during its expansion and contraction phases are normally distributed, instead of deriving the distribution at changes in the eLi directly from historical data . Finally, in calcu lating the probability that the current eLi observation is signaling a recession, if the Neftci formula said that the probability of a recession last month was greater than 95 percent, we set last month's probability equal to 95 percent. This modification is meant to prevent the probability of a recession becoming stuck at unity. Once it has been determined that the economy is in recession, the Neftci formula can be used to yield the probability that an expansion is imminent . This is done by switching F~ and F~ in the formula and replacing PL with the a priori probability that the eLi has entered its expansion phase, given that a month earlier the eLi was in its contraction phase, :!l1:tI~ . . .1 I ry l1l ~ 10 !on·cl .... t n:~~· .... _ '1(>n.... :l1ld l.·\p.lll- reCl.'nt 11l:l\II1Hllll. Of h~ l()olk ill).:.11 tlw g:IP ... ion . . Ill'l\\l'CIl Ilil' 1H1I1ll'rO ll . . . f~·\·i~i()Il ..... :1I1d l ilt' dmk\lh~-in I'l':d nI ITl.·nl \ , Il"l' (11 111~' ( 1.1 :!nd ;1 11 1()\·in ).: :1\ ('I': lgv (If rn ..: 1l1 I : dll l·~. IIl:lIl hy looking ti!1l~'_l t kn()\\in ).: \\Ill'lh~r 1'111':1 l k:~·: H I~l.· 01 :l illll' 1110n l h jlllblk:llitlll (k by. til L' ;nd ...·x h:l . . ur 11a . . not R';lchnl a i.: ~cllcd turni n g pnil11. 11K' eu can. :1 1 11<.: . . 1. pro\'idc ;1 ('oinu(k m . . ,gl1a l o f ren: ..... io n .~ :md (.." Ix,n .... ,on . . E\('n then. tlll'n.: .In: likl.·lr 10 b ... lll:tn~ 111onlh .... 1[1 \\ h idl 11 1<: CI.I . . ,W1ab th.lt . 1 \,:11 1l1lm1K'r Oll·(lIh",·nlt i l l.· Ikl'iinL' .. . in lh .... TIll'l'v do ...'~ n ol .tpp...·.11' 10 lw :l11y ;1(1\':111t ; , ).!~,. 111 p r .... dll: lln).! I'.... (·l·~, i (ln ... 10 1110\ Ill).! from . . il11pll: ruil· .... 111 I Il1llnh to 1ll01'\' !'Iopll l . . tic:ll ed ....igna li ng 111l'1I11x.1( )lt»).:Il·~ fl·U· ......... ion or cX lun, jun i . . (k\l·lopl11,1.! \\h ... l1. 111 !:tel. ont' dol.·' [lut l11al~'ri:di/l' Iluk .... of thullIb thai do Ih l · IK'~I loh 111 prl.'tlidm).! fl."t'l.· ........ tO!l .... appl.':l r 10 lx' Iho.x · tlut piau' :1. . Illlldll·lllph,l .... t .. on I II(' ,/ ~('ol dl'cll1ll.'~ III t Ill' eLI : ..... 011 till.' IlIIlIIht'I'of d .... dinl· . . Thll ..... olle :tp p:l rl'nlly dOl,:' 1x.: II(·1' h~' look11l}{:11 Ihe diITl'll.·nCl·1x... I\\l...... n till.' ItlrlVllt l;llul'I" tlw C LI ;lnd;1 g il il1d~'\ I II prl·ll illlll).: l·\P,IIl .... !Cln ..... III l'(mtl:l .... t. :III (II' nlll· .... til' l h umh pl... "f,)fIll ;11)( JIll ... qu:lll~ \\dl. hu t 11K' 11101'...· 'Opill"liCIlL'd '\dtci IHe l11lxlolo;.:y h;l~ :1 cI ....,l f .Ilh ;tlll;[gl' :-....·1 n :lll-alltiol" .Ifl· i n Ord(T Fir....' . Ihe t~ 1)<.' .... ;111:11\ .... , .... i ll 1111 ..... Irode '" h:lx'd 011;1 Il'n lilllitl.'(.1 "1I11pll.' \\ l' h.II ...· l.'\I'Ii,.. nl.·ll1,'l.'d on ly li\ ...· 1'l.·cl.""on . . . . mH· till' lI1Il11)(l'i ll' IIld,,\ of k:ld1llJ.t IIldlGllllr .... Is an End to the Current Recession in Sight? The Nettci probabilities of recession and expansion plotted in Figure 10 incorporate values of the composite leading indicators through March 1991 , The figure shows the probability that an economic recovery is about to begin rose from 2.4 percent in January this year to 15.3 percent in February and to 30.3 percent in March. The April value of the composite leading indicators has now been released (along with revisions of eLi values for November 1990 through March 1991 ). Applying the Nettci formula to this new observation gives us a 56.2-percent probability that the eLi has entered its expansion phase . On the basis of the historical record over the past four recessions, as summarized in Figure 13, this probability suggests that an economic recovery may already have begun . However, the probability that the eLi has entered its expansion phase is still low enough that little confidence can be placed in this prediction. \\:1 .... fIr . . ! dL"n.'lopt.:d :Ind (;11 til ...· !Ill1\." thb al1idL" (ml~ hlur rL"{'i)\ ...'ri ...·. . TilL" :Kh :lllIa.o;:...· \\ .1 . . \\ rill ...·n) Ih:11 one ruk of tlllllllh .IPP:I1·...·mly 1Ia .... tl\'l'r :lIlolh...·r lIIay dj...,:IPI'"H:;lf or 1Il;1~ Ix' n;\·L"r ........d as :lddiliolUI o lher\·alion ..... llTl][)Hll.lIl· &-cond. IhL'I"L' h:l\l· I ~.'L"n IUllIIL'f(llI . . {ldlnili()Iu l dunJ.!L·.... lo II1\." l ·Ollll'"HI . . 't ...· intl...·x Ihrmlgh Ihe yL·.lr ..... "'>(1 Ih.lI IllL' II1d ...·'·. . 11i . . !<)rk·al 1>t..'rf()fIll:lnL"C i:-. nlll \1\."(L· .... -,;lri ly a ,W\(KI gllilk 10 t i ll' flllurL" bch:I\'ior of 11lL' inck'x liS (.//nt'IIII\ · d(:fill('d A lI ....l·flll cxcni . . l'---lx·yond IhL" .. .n l p ...· of Ihi .... .l1"llck-\\t)Uld 1)\.' III 0 .......· rl'o,I-lill1(' dala on IhL' lIld!\ idllall..·o111ponL·I1I . . 01 I I)(b~ :-. <..:1.1 [0 (kll'fllllllL" ho\\ l(j(b,,·. . CU \\lIuld ha \'e heh:l\·L"d. 1h rollgh li m ...... 1.... d:tI:1 o n Ihl' ind i\ idual l..i)lllPIII1\."rll" \\'t'rL' i n ll i;llly rl..·k·::"l·(! . Fln:III~·. \\L' r"-'pl·.tt [hal 110 m:lllL'r ho\\ liuk' " ":! rning of hll." ' nl· ........ cyell.:.· lurning poin[.; IhL' (11I1I 111I..·rCL· I kP:lrlllll·lll'." le: ldillg i n diC:llO "~ "L·ril..·. . . m:ly gin:. i[ 1ll:I Y .~ l ilJ p rovidL: IIIL' L':lrlic,,[ \\";l rn ing :1\':li bhk F.conumic KC"ic w - Ju ly 199 1 13 References Dit!bold, F••mcis x., and Glenn D. Rudc.::husch (99 1), ~Turning Point Prediction with the Composite Leading Index ,~ in Leadillg £CoIlomie /Ildiealors: New Approaeb£'S and Forecasting Ri.>c:ords, t.-o. Kajal Lahiri and Geoffrey H . Moore (Camhridge: Cambridge University Press), 231-56. - --, and - - - (990), "A Nonpar::unetric Investigation of Duration Dependence in the American Busint!ss Cycle," jOll n/al of Polilica l Economy98 ( June): 596- 616. - - - , and - - - (1989), '·Scoring the Leading Indicators," '/011/.,1(1/ of Hllsiness 62 (July): 369-91. Hymans, Saul H. ( 1973), ~On the Use of Lead ing Indic:ltors to Predict Cyclical Turning Points." Brookillgs Papers Oil Economic Aelfvily, no 2: 339-75. Mankiw, Gregory N., and Manhew D. Sh:1piro (986), ~ I\'ews or Noise: An An:llysis of Gl\T " Revisions," SIITVey ofCllmml Bllsiness 66 (May): 20-25. Moore, Gt.'Offrcy II ., and Victor Zarnowitz (1982), ~ St.'quent ia l Sign:lls of Recession and Recovery," joun/al of Business;; (January): 57- B5. :>:cftci. Salih N. (982), ~Optimal Prediction of Cyclical Downturns,' JOIln/al of &Ollomic Dynamics and Colllroi" (November): 22;- 41. Niemira. Mich:leI P . (1991), ·'An International Application of Neftci's Probability Approach for Signalling Growth Rccessions and Recoveries Using Turning Point Ind icators, ~ in Leading Economic /lldicalO l"S: New Approaches (lnd Fonxaslillg R(.'Cords, ed. Ka jal Whiri and Geoffrey II. Moore (Cambridge: Cambridge University Press), 91- 108. PhillipS, Keith R. (990), ~The Texas Index of Leading Economic Indk":ltors: A Revision and Further Eva l u:llion ,~ F(.>dcrn l Reserve Bank of Dallas /.:Col/olllie Review, July , 17- 25. Fedcntl Resc:rve Bank of Oallas Kevin J. Yeats Associate Economist Federal Reserve Bank 01 Dallas AReturn to Profitability: The Performance of Eleventh District Commercial Banks I n IhL' Ek-wnth F\:.·<.It.·I:ll RL'''>L'IYe Ohtrin. J()l)() \\":1.<; :1 turnarou nd YL': tf for tlK" comnH:'fl'ia l h:mk ing industfY . ('I"Ilt' ElcvL'nlll Di:-.trict L'nc()lnpaSSL·.~ ·1·L'xas. nOll11t'rn In ui.<;i:m: l. and :-'(lutilL'rn :\'L'W l-.1cxico) Hanks in IhL' Dislrict colJectivdy ea nwd:1 pmfit for till' ti l'sl linll' SiIKL' 19H,). :lIld IItL' share of profil:lhle ha nk .. rOsL' from ')6 l'lIo: rt"l;~ nl in 19K7 10 79 Ix'IH'nl OntO b l"lor l"onlriiJulin,l.! 10 I IlL' illlpro\"t."d I")o.."rformann' \\a .~ a si,l.!nifiGmll\::du(1ion of nonlx'ditnn ing lo:m:-. :11 o JIIIIIIL'rdal hank:-. iliad... possihlc hy a sl L"ad~ fl'(U\L'I)' in Ihe rq.:ion:11 t"('o nolll} .. mcL' pm7 :lIld till' n.....olulion of the.: most trtH.lhk·d IXlIlk.... hy till' Ft '<.k·r:.lllx."po..... il Insllr:.Hlce C.orpor:.u ion ( FDIC) I I n Illi. ..lI1 ide, I (."".\:II11 int' tht' condition and lx.-rfonna IKt· of l"ollullerci:ll hank~ in thi,.' EIL-\ L'nlh Ft·tIL-ral Ik~'r.· t, Di:-.trkt in 199() 1l."'\plorL' Ih~." ft'a~()IlS Ix.-hllld Ihl." ha n k ..; impro\"t'd pL'rfolll1:I IKe [ :11.">(1 L'x;uni n(' Ihe rok llIi,.' ['l."ginn ·s t"conolllk fl'CO\ L'I)' playi,.'<.[ in h :mk pl."rforlllant't' imk-pentlt'lll of till." dl"t'\l o f IIlL' FDI C .. resol utioll of insol\'L'nl h:lllk ... ;\ ly analysb ind icak·.. IIt:lt bOlh l ilt· a'ginn ·.. l."cono m ic improvemenl and Ihe FD IC a ..sist:IIKt' ht'lp<.."(l red uct' nonperforill ing loans :1( ("\II11I11L'I1.'i:11 h :lI,ks. thus contnbut ing to net prolits To dt'tL"nnin...: how h ro:ldl y hased thL" I\.'CO\·CIY h:I.. iX.·l."ll. I di\'ide hank,.. ill\o Illrel:' sizl." c ltt·goric.. :lIld look :It 11K" rx:rfonnam'e of e.:ach group I ;Il."tl Olll1P;II1.· 1'.lL'\cmh ])i.~lri("\ hanks \\'itil lXlIlk<; in Ihl: rest of l ilt' lUUml) I find th:lt b eg.... hUlks III l lIi,.· S<lulh\\ t' .. 1 11:]\ L" nearly {':llIgill up 10 Iht" p,,-,rtorm:l!lct· k\·d .. !(lund d~'\\lll."l\:: in Ihl." {"()Unl ry. llUI small :tnd 11Ild - ~ i /.L' h:tnk .. han~ room for impIU\'t'II1L'm Assess ing the performance and condition of Elevemh District banks Thl." hanb' rX:rfOrtll ;IIKe lI~tn~bled into a return o n ave l~ lge :IIooSt"l .. mOA) o f O.</j pt:ft"\:.'nl , sl i,l.!htly It:..s than th t: 0.':;0 pt:rcent repol1ed hy banks oUl side tht: ])b!riLI.~ lklllrn o r' :IVt'r:.t,l.:t· equilY also showe.:d ~:tin .. :tnd . in faCl , W;IS gl"i,.':tIL'r for EkvL'mlt Dis!rit.·! h:tn ks Ihan for 1>:lnl.;.<; d~ew lK"rt·. rt"lk:cl lng lltt' lowcr capit:111~lti()s rl."portl."d by EIt-\'i,.·nI Il Dislrit"1 hanks (S('(' fbe I)(A\'/illed "/)(11I1ilili liS (!I" Il(/Ilkill~ Terms " ) profils also incn:;L ,~t."(l :I ~ Cfl!dil quality irnpro\'t'd \,\' ilh bctter cft"(fil qllalit y. morl! 10;ll)s p:lid inlt'rt':-'\ .lTld intcre..,1 in(:omt' rosc. The rb k pr\.'lllilllll abo fdL di,.'Cr\:.':hi ng tht' intt'rcsl Olst o f dq)osih ' On the hanks' incomt· st;tt t'IllCn\S. lnwl."r premiuLlls t"atl~""(1 tOl af interest c'\J"k.'nM:' to <It'dine , ~lOlm"lQ ioans. 01 de IflQuonc es. are delJned as loans !hat arc 90 01 more days paSI due and still acCrUIng inlefest, plus loans If} nonaccrU81 StBtuS Those ROA f'gures for 1990 Bra doM! shgntly from !he h,storocat tl vcrBge o f O 5€io;Cfcenl rceordod over!ho t985 89 ocrooo for all banks nm'O/lw'(!o Th,s Ioguro 's blasod downwilrd by 1Il6 {XXJr per lormance 01banks on rhe SOulh, west Ihe f'gure also includes 1987 a yesr m whOCh many large OIJnks rcpotrod SUt)Sumloal/osses relolfJd IO i00ns 10 dOilelopmg na llO/!S Oeposo/ors WIth balances exceed'ng the $ lOO.DlXJ depoSll HlSlK8flCC limit want a hIgfIcr !I!lcrcst rOlC to compensate for the ,,$I( 01 lOSS of the bank farls rhe grealer tile Chance 01 fadure the grea ter /he fIsk Plen'IIaTI dcpo5J/Ots demand (Snort and Gunl/lerJ The,,$1( PI~ had /OOeased as depos,tOtS demarIdccJ a higher IIItcrest rate r/erJOSl/S such as on t¥W1SJ.JtOO Iar~ c(!(/,flCares 01 depoSI/ The ,,$I( pretTIIUfTI paod by EJOvenIll Dos/rlCl banks has declined over Th ~' 13{)~ nlmmL'R'i.11 h:lnk~ in Ih.: Ek'\"(:nl h Di:o.lrit·! L':II"I\t"d a ("Ollc(11\'e S ~K3 mi llio n in 1990 ".cono m lt' Rt'Vic ... - Jul)" 19') I the past several yeatS r~al'VC to the rates paId by banl<s ~ tho resr 01 tho COlIflrty " (ro lll 52.'1 hillion during 1989 [0 50.9 hillion in 1990. This Sl 5 ..hil1ion decline in cxpc:nS<.,:, k'd Ihl.' w:ly for an int'rt::I'ot: in nel income of 5 1.:\ h illion, The :111l0Unt of !nan ch:lrge-ofl'.. . dl.'clinL'(1 h}' $750 million !X'1w,,:t:n 1989 and 1990 Tahk I pf()"kles tlK)l'e dctai] :IIJOlJl Ihe cum.:nl and hbloric:l1 (,·onc.lilion :md pcrfonn:ul(;c of Ele\'cnth Di."' rl<.1 hanks. Figure 1 Total Nonperforming Loans, 1987-90 " Factors contributing to the demise and recovery of Eleve nth District banks "" ... , '988 "" SOURCE OF PRIMARY OATA ComoliOaIEtd Repons oj Condition and Il'ICOme (Board 01 Governors , Federal Re_ Syslem) and net iO(.'om e to ri ..;c O ther effect1> of better nt.xlit qualit y also il1lpro\'(:d profitabilit y A'1> r:n::dit quality impro\'ed, loan.. . fL'quifL'd less monitoring and b:Ulb p:lid few..:r fI..'cove!')' COStS, thus nonintercst eXI)CllsL's declined. Ilaving more perfonninl-: 10:11110 and :1 lower cost of fu nd... :llso nlnt rihulL'd 10 Ihe in<: reaM! in the hanks' nel intt:rt:.. . t ma rl.dn U)WL'r market in terest rates helped improvL' h:lIlks' profits The net imere... [ m:lrgin generally impn)\'c'1> :1'1> inh.:rcsl raIL'S d ecl ine, which they have donI: si n(.'e carly 1989. The boost from IO\Vt,T intL·re.. . t r:.ltes tl...;tl:llI)' ( X't'l If"1> when deposit rales fall fasler Ih:1n knding l~lIeS Eleventh Di'1>triCI hank profits rose in 1990 because of both innL';lsnl revenues :lncl 10wL' r expenses, IntL'rest income grew by 5 14) million in 1990 ovcr 1989. CVL'n as 101:lllo;lI1S dec.·lined lbnb no lon).:er h:ld to make !:Irge additions to 10:111 loss n:scl'\'C1> 1'lt:c.m:.C' no nperfo rming 10:lI1s had c.k""t'f'c :IS<.""t1. Additions to the rt.""1>t:J'\'e account fell Nonpcr/otfJ1j()Q loans reached a peak 01 S 10 5 biIhon IIllhe Itld'cJ Qvarf6' 011988 ,6 Elt.'\'cnth l)iSlric.1 hank·; rt: itlrIl to profitahilit y in 1<)'-)0 s\t::mnlL'd frolll three f:Ktol"S , First . Ihe h urdt'll of nonpcrfo nning loa n .. suhsidcd as h:mks realized their lossc... :tnc.l relllm'L'd ]);Id loan...; from their hooks Second, Ihrml).:h thc rc.~()llIti(l n o f f:lilL'd hanks, lil(' FDIC took over m:lny of the nOllpcrforllling loa n!'>. Third , with the nxovefyof til... rcgional L'conomy, som\! borrowers who Wl.TL' delinqut'nt in their loan IXlymcnts werc ahle to IX'come curren! o r to rcstmcl urc thei r IO:ln..; and llK...·\ n\::w paymelll sch<.-"duk1>. T otal nonpl:rfonnin).: loans al YC:lr-t:nd 1990 were ic:;...; th:m onc-tliird thL' It-vel rca c1l1::d ;11 [he end of 19H7 (FiR/lit' f) I Fewer nonpcrfo rming loan l osses. U 1SSC'1> from d<,'linqul'nt loam piaguc<1 Elt::venth Di.. . trict hanks in thc 19H<}.., :l1\d. in too m an y cases, resulted in hank f:tilttre:-. (Sec' ,he box filled 'Hull' &mks Copt' wUb NUl1pelfonllil1JS UxIllS ") N() nperformin~ loans, :l lthou~h a 110nnal, exp(xtec.! p art of h:mkinR. jeop:lrdize ;1 b,lOk 's wdl .. bdng when they :Ire we:ul..'r th:rn lmt kipatL'(I, which W:IS tht: C,';1..... e in the Elevenl h Distri ct in thL' mid- to laIC 19HOs. By underestimating nonpcrfonning loans. many Eleventh Di~lrkt hanks suffered losscs that diminisllL'd or depicted their 10;ln 10$." resel'\'e and equity c:lpilal accounts. By Ihe end of 1990, Il()\\'evcr, mosl Ek:V(;llth Districi COlllmercial hanks h:l(\ been reliewd of the burden of nonperforming loans FDIC resolution, The FDIC's method s of resolv .. Ing failed hanks. in effect. n."'(iucc thc amount of nonperfolllli ng IO:lIls rcpt: )rtt.x l h y opcr:lling ('ollllnercial hanks. When a hank 's cIJaT1t:ring :Iuthority declares the bank insol\'cnt , the FDIC UML:llIy bt..-t-,ome:-. the bank'!> receiver, :md ;Ill the hank's :1.....'-t1.~ :m d liabilities arc transferred to the FDI C for disposition. The FDI C has two h:lsk resolution methods at its di5JX>Sl11 : purchase and as...;umption tr.msaclions :mel liqllidations \Vith Feder.1i Rncn'e Bank of DaUru; Table 1 Measures of Financial Condition and Performance OIh'" Districts' Saoks All Elevenlh Dislricl Banks Year 1987 1988 198' 1990 1990 Number 01 Banks 1,903 1,624 1,437 1,305 11 ,007 201 .2 110.1 11.0 3.' 2A " 160.4 90.8 58 8.7 3.2 2.6 163.4 84.2 SA 87 2.8 2.0 180,6 80.6 2.5 11 .1 2.2 1.2 3,166.4 1,960.7 75.0 206.4 52.6 27.8 3A - 2.1 28 - 2.1 2.' -.5 .8 30.7 15.7 (In billiOns 01 dollars) Talai Assets TOlal Loans TOlal Nonperforming Loans TOlal Equity Capital Total Reserves TOlal Net Charge·olls TOlal ProvISions lor Loan Losses TOlal Net Income (Percent ) Nooper1ormmg Loan Ra1io Troubled Asset RatIO Net Charge·off RatIO Loan loss PrOVISIOns 10 Total loans Pnmary Capital RatIO Return on Average Assets Return on Average EqUity Net Interest Margin ., 8.57 6.47 2.08 6.35 5.05 261 6.45 5.05 2.24 314 257 145 3,83 2.96 1.43 2.91 7.29 - 1.02 -17.57 2.76 2.79 6.49 -1.08 - 20 ,94 2.62 2.78 6.14 -.30 -6.25 2.66 112 7.27 .43 7.94 3,01 1.59 8.00 .50 7.84 361 Percentage of Profitable Barils 56 61 69 79 Number of Failed Banks 67 121 14' 105 ( Percent) Failure Rate 3 6 , 7 . 64 m)\1 pert()nlltn~ IIKIII~ ,Ind ()[iler [Hlllilk-d :I ~"l·t " !i(!lIid(/litlll, thl.· FD IC f<..'Lli n .. :111 till' Cl ik d h:II1],:,> :I ......C[ ... 1)( ,tb ~\ )()(I :Ind Iud. ,tnd d i.. p( I'C" • ,I' ,IK'III from thl.' h;tnk thr{) lI ~h I.:llhl.'r llll.'tl1od , till.' l" [)lC rCI11u\ l.'~ till' [ n ,I /1/11"1.1>(/,(, IIlId (/~Sllll//lli(," d li:c[ of I" [)i\l~io n 01 LI<.[lIId:l! ion ' 0 h .l n], tnlll,' f/( 1/(11/. \\ hit Il i~ ll"\'\1 III Illll.. t l);Ink LlIlul"l.' l':bl."'. ,I pr~' ,lrl';tn;.:\: d t,l"l.'~ P()~-.I.'""i(Jf) (lr I I ll.' )..:()()(I .l'~l.'b 1~ljk-d Illl\l.'r frolll t ill.' Elil('d 11:111],' trouhk-d 11( )l1 p t.: rfnrlll l11),! [o;tn~ [,;1 .. \ ' of rl'\ h,m" TIll.> I 1)[<. ;t ....l.·b. IVI~lin .. till.' indudlllg thl.' '1h l.' h UYl'r :K<jUII"l'" .1 h nud ,'nul.',).:.\·nl'r:lling ;I".t.'[~, ,ltll" illl p ro\ in;.: t lll'llll~l'r'" ()\\11 PUI"Pt.Th Ilu pf()fh:ll,ili[\" In :1 I..Jn<Jcr .II ,,-/loll:' /JatIJo. (JUfchaS(l.l"d <J55a>'plu, IIl9 acqtJiI . mgoanlo.takesallll>!!assersotmcl.'l 'edb.l/JI< ~JClud"'9a/ly nonoer1om'lIlg b1n$ IN:1/oteclOSC(l ,cal erua'f! TlIf' FDIC compensare5 /tleacQlJfftngo.ri Dyptov1(11J1g II 'M/h(!fJOVgII C./sn ro cha'ge 011 me loonperiOt""rg 1o.11IS 17 Definitions of Banking Terms Banks Failures include banks receiving FDIC assistance 10 prevent failure. total loans 90 days or more paSI due and still accruing, plus nonaccrualloans to total loans. Failure Rate is the number of banks thai tail during a year divided by Ihe number of banks at the beginning of the year. Primary Capital Ratio is ratio of the sum of equity capital plus loan loss reserves to the sum of total assets plus loan loss reserves. Loan Loss Provisions to Tot al Loans are the total additions to loan loss reserves from current-period net income to average loans" Return on Average Assets is the ratio of total net income earned to average assets . Net Charge.off Ratio is the ratio of chargeoffs on loans net 01 recoveries to average loans . Net Interest Margin is the difference between interest income and interest expense, adjusted for interest income exempt from federal taxes. divided by total assets. Nonperforming Loan Ratio is the ratio of dirL"(:liy :.cqllin.·.. 11K" Llik'd h:IIlk":-. nllllpt.'rforming loan:-. ill :1 liqui(i;llioll. Fro m l~nC Ih roll~h P),X). 1."- LOIllIlK"rd.. l h:tIlb . failL"d ill Ih~ Ek\elli h Di"l nU Tho.:."L' f,l ikd htllb hdd 10lal a:-':-'L'h of "':-;1 (I hillion. or .Y-l p~r C~!l l of Ilw tot:l! .1'0,,,,,-'1 .. in 1)1....Kt hanl.;:-':11 1110.: L'l1d o f I l)~(). Of 11,0.:... 0.: Ij - failul"l':-'. 10 I \\"~f'" re:-oln'd hy I h~ FDIC by sdlitl.1..t 11I~ b:tn~ing OPL' I~l l i()n.~ 10 o i lier hank:--., Only 1(, W":I"l' liquid:IIL'd. \\"ith 1110.: in.,urL'd dqlo:-.itor'i reo..'idng p; l ~ l11~tlh from (Ill' FDIC. Th ..., "L'l11ai n in~ 17 \Yer~ gi\L'1l Open l\:tnk Assi.' lam:..:, IIlL";Hlil1).: Ih:lt tlw h :lnh \\:1, not I(,nnally dO:-.L'd. hut Ih...· FDIC :tnd olll,' ltk' tn\ L',lor' pro\itkd mor..: c:,pll:.1 \\ h ilt' Ih"" h:II1" lO11l11ll1...'(1 ih 0fX"t:l1iOIl .... UJ1l1ll1...·r<.·ial hill'" 11,1\0..' .11,,1 ho.:ndilo.:d frolll Ic .. ~ ",·olllp•.:tiIJOIl for tkPo:-'lh 'tnt'...· th ..., Ik ...olulion Tm:-ot Corpor.ll ion IItTC1ICM)h control of or .~old Ill:llly elf Ill'" ill., ,)" (:111 Ilml1' I h,1I \\ L'fL' Ol"k"r.nillg anus., Ill\.' l lnilL'<.l ~:II ...". p:1I11("1I1.IrI~ in th ...· SOllth\\ ....... 1. Tht:""" lIl'-tlh ...·l1t thnlh p:lid ""\(""" ," "I\dy high " Return on Average Equity is the ratio of total net income to average equity. Total Loans are calculated as loans net of loan loss reserves and allocated transfer risk. Troubled Asset Ratio is the fraction of total assets that are recorded as nonperforming loans or other real estate owned. Foreclosed real estate assets are recorded on the bank's balance sheet as other real estate owned . inl~ro..·'t 1~llo.:" to :1IIr.1...1 .ll1d "~I:tin dqXlSlb (Shon ,Int! t;lllll hL"rl . .'!in.......· thl: d~ lIl i.....: 01 Iho.: irNlh l'nt lhnl1~. Ilmlis ;lnd \"( 1111111",'1\.1,.1 hallt.., ,lliko..· lu n: Il\.·nd"it..:d from nol 1l:1\ in).: to 1II.lIel1 high dt'po~it 1:11 ...·'. In dli.:{1. [hL' Ir['(:', n n1' ...·I....:llor,llIp :Illd ro.:.~e llu[I(J!l ;tnh it~ 11:1" .... Ie Ip pt:d tll11~'!1 of [ht;'" "hidd ill).: up" of d,,:posl\ IIlIL·I....... \ I~IIL'~ h y trollhl..:d l hrirt.~ . Tltroll;...:h tilL" ..:nd 01 1')lX I. 110 thrift~ in Ih...' Ek" L'n lll F...·tk-I~ l l Ik:-...:n"L· nt.,triet h.ld 110..:""11 ",,-·il.L'd hy th ...· HTC. Of til ...·.....· ~"'il.tI .....·... H I t ll nt1~ \\ ...·t..... 'iold to IlL"W 111\'...',10)"'. tYPK:llly (Jlht:r fill. )I\tt.l l ill'tltlltion~. \\'110 :I ....quiro..'(! Ih ...· Ihrif(, l"M..'rforlllill)..: :........... t ~ Regiona l econom ic recovcry, In 1110.: Ek'\l'nlI 1 I )i~1 riL"!. ]I)'X) W,I, ,I y...·:lr (Ir ~n 1lllllllit' o.:X IXllb!t)ll. Tt ..... up,,\\in).! ...·I1.thk·d 'tlill<..' t!...·hllquL·nl hOITO\\,"-·'" to h",·\..tlllw l:lIrr",'nl III Ih~lr 1()~ln p.I~·mL'llb 10 h,IIl " ' . .1Ilt! It :tl IO\\L'd (·Clll'llll· ...... I.1 1 h.m"'. IIllllrll. ttl ro..·tlttl: ..." 111""ir Il(JIlIX'l"ltll"llllll)..: t(I:lIl' Empl()}"I ll"'lll daU pl"ll\id.... ",·\ ido.:rKC Ill.1l1111h...· IJL·\..:IlI11 Di~lIit'l Ix.·).!;tn to I"...'('o\",·r hum ih regional h-..kr.. 1 Jh . .. (·I""\·c ha nk or I)a llas r\.'Cl:,":-.ron of th",' btl: 19xos, I':ryroll erllpluynwl1t has rnOl"l' th.m rt."l:o\'l'rl'd from thl' ~)h lo:o..'ie." experiI..,nt'l"(l in tht, do\\ mum of 19X(, :md t>:H~y 1987, Emp!o}'rlH:.'nt has ;-oho\\'11 ,,II..'ady ~ro\\'th, with ;lnnual irll.:rl..';I;-.I.::-' (>f nl()~ th:tn 1')(1,000 ~1I):-. fr(Hll 1987 throu~h 1989 " In 199(1. ho\\ l'WI', ~ro\\'th :-.lmn,"(1 til ;t -.tOOO-jclll i[1/:I'I..';I'\I..·. 'nIl: annu:rl a\'I..'r.1ge llnemploymcllI ratl! i" down nm:-.itk'r.lhly from that l:Xlx·t;l·tKCd throu).:h IIl1..' rllld - J9HQ:.. in Ihl'l..-,(;, of lilt: Di"tri(1:-. 1lla~)r melropolit:rn ;tl'\.';I:-.: O:rll;r.s, Fon Wrll1l}--r\rlin~10n, :11)(1 IhlllSlrU) The banking sector's delayed response to the region 's recovery Tlrl' E1t'n'nth D i:-.lrKt hanking SIX'lor's t't:cO\'elY follOWL'd till' gl'rlL'l:d l't.:ono11lk n;.'co\"t:'.Y with :1Il app:lI"l.'nt bg (If thr"l.'1..' 10 four Yl::I1':-.. TIll..' bg may lra\·1..' been esrll.:t:iall y long IX'C:I\l.~e of tht' depth of thl..' n..·('e....~i()n that hit thl' SoUtl1\\'l'.~t. 1\,'; region:l! I..·l,:onorniL' dO\\'lltlI1'l1:-. tlm::ltl:n other area:-. of thl..' ('mlnlry, hmkl:'r':-. m:l)' IlI.:ndlt fmrn lInderst:lIlding \\ h.lt ;-.IJmc of till' Smlth\\ l'SIl:rn h:rnk;-o did Ihal ('n;rllll"(! tht'lll to ;-our\'i\'l' 0111.' of tht> Il)()t,t important k-....·iOtb 10 111:: 1t':UTIl'(l from thl.'''''-' h:mk~ i~ to not umk·ro... .,till1al~· lo.:.rn In...... rl'''''-'r. I.' fl.."quin:"Illt'nl';. The recove ry o f unass isted ba nks_ To 1e;1rr1 how fa:-.t the South"l·,.,!'., improvcd lXonorny tl~Ul:-lall-'d into impron:d ("(mullerci;r! hank perrorm:IIl{'C, I cl':al11int.'(1 the I:onduion of a ~roup of h;lIlb that \\'e~ inI1Ul:IKt.'(1 only by the rl..1.:0\'t.Ting t.·uJIlomy :lIld nxei\'l'(l no FDIC :ts~istam:e Thl'SI.' UIl:I~.,;"tl'(l h:Il1k." nt.'il lrer :K(jllil'l.."(I:1 failcd b:lIlk nor bikd thel1l:-eh'I..'.., in :1 :-lIh:-.e(jut:nt ),e:lr; thcy 11I:1Il:lgt'd 11lt.'if nonpcffonnillg Jo:rn .~ without help from IlrL' FDIC COIhequelltly, :lny irnproYerncllt in 1I1L'if ~'(J!ldili(Jn l'dkl'll'{1 primarily Ihe illlpn)\'ni ~'~'lJll(Jm y 1\ly l'x:lIllin:llioll l'evl':l kd th:1I the eOlllolllY i lllluelKl's h:rnk Pl'riol'lll:IIKe wilh :1 deb)' of a.~ long:rs three.: Yl'ars. Till..' EIl'vr.:nth Di.~trict l:conomy :-.ho\\ ed Si,\!lh of rn'o\'l:ry as e:lrly as 19ie ~t,\,t'l1hde:-,. thl' nonperfOnlling Joan r.llio :11 Ihl'''C lIna,,:-i"led h:II1"., dillllx'd to nearly (, IX'R'l'nt in 19H9, B~' 11K' I..'lId of 1m. howcwr. thi" r.rtio bad fallt'll to Ie" than I pcn:ent, C:lpil:rl r.!lio:- in l<)t)() :rho Il1IprO\l'(l from 19H9. While rl:1Urn on ,1:0..",,-,'" (~IIlIl()t hI,: de.,nilx"'(! as rol)[l:-'1. b;lIlb m:IIl:lgl'{! to ke~'p l(lv,I.::- during this pcrio<.i at moder.rte 11:'\1.'1'> A hl..·althy net interc:-;I margin Ec;:onomic K.,vkw-Jol), t 99 t Figure 2 The Recovery of Unassisted Banks, 1987- 90 ....... • , -- • , 1-_______-----"..--.-.~.....~'IIIwn.• SOURCE OF PRIMARY DATA. Consohdaled Reports 01 COndition and Income (Board 01 Governors, Federal Res8Ne System) I..·adl year helped limit net lo:-.,c:- (Fi~/UY! 2) This group of hank:- :-till ha., room for il11pmvcmcnl. Thrl·I..' Y<":rrs d:rpM.'(l 111::lwtc'Cn the fir:-.t :-i).!n~ of t'conomic recovery and the hank,.,' return to profitability. The Glpit:rl r:llio~ of tlll':-l' h;lIlk:-. willt;lke e\'l:r) morl..' lillie 10 fully reCO\'t'r from the l"Conornil' downlurn A::. Ihe economy in Ihe E!en::nth Di.~tri<:1 cOllliIlUt.·., 10 improve and as thl..'M:' hank ... gener.llr.: profit. they also will he abk to inneasc capit:11 and dl:lr),:L' off more of Iheir nonpr.:rfol'rning loan:-. Tahk- 2 provides more information ahoul thi.., group 0 1 hanks. A lesson for other banks, Tlrl: Southwest hanks' ('Ou r:-.e to rec(J\'cl), I'I..'\'\..':I)S :111 important Jesson for finand:11 inslillllion,~ ,~ufkring llmJugh region:11 l'('(lllr)Jllic downtllrtlS A)I EIL'vl'nlh District hanks opef:lled in Ihl' ~all1e l:l..'onolllic environment. hul not :Ill of them \\'1:'1't: atl'et'lcd 10 the s:lIne degree, Some b:rnk., wcre ;Ihle 10 ('onlinue operations wilhOlrl oUI:-.ide :I ,~:-.i.~ta .Ke \Vh:11 the su('<..'essfuJ h;rnk" did for their o\\'n Mlf\'iv:d was 10 :I('cur.ncly fO!'t..'{·a,.,r the depth of t!1l' fl'('e:-.~ion and pn::p:rrt' • EmpIoyr'nenI I'9U'es arc /rom seasonally adJUSred empIoy- menfc1<Jla 19 Table 2 Measures of Financial Condition and Performance Unassisted Eleventh District Banks UnasSISled Banks Year ,,.7 1988 '''' 1990 Number of Banks 1,443 1,279 1,181 1,120 97.4 48.5 81.6 79.4 35.7 (In billions of dollars) Talai Assets Total Loans Total Nonperforming Loans Total Equity Capital Total ReseNes Total Net Charge·oNs Total Provisions for Loan Losses Total Nel Income (P8rcent) Nonperforming Loan Ralio Troubled Asset Ratio Net Charg&-etl Ratio loan Loss Provisions to Total Loans Pnmary Capital Ratio Return on Average Assets Return on Average Equity Net Interest Margin Percentage of Profitable Banks for it:-. il11p:I\'\' 10 d,.:n.:lop :1I,d m:H1lI:lin :I mOl"\" di\"cl":-.ifi ..:d portfolio wilh limilcd l'xpo:-.ur;.' 10 Ih;.· \\ '.:a b.:r .~l"("l()r.~ of I hl' (k-.,:I in ing SOUlln\ <,.':-.1 cconomy. ancllo m:lk ...· .~O\ lnd k'nding dL"("isi()n .~ Assessing performance by bank-size classifications: h;lS the recovery been even? To tIL-lcnninl' it 11K' h:1I1king n.:n 1\ cry h;l.~ b...·\:n broadly ba:-'l'd, I :-'l'pal~lll'd llll' :-.;unpk of hank:-. illln 11m.'\.' group.' h:I:-'l'd Oil :-.ii'oc D IViding IIlI' h:lnk~ hy ~ii'ol' fadli[:tlcd ('olllp:lri",on, \\llh pn:\'iou:-; ",lUdic' 111l" d"lim:ll(>n 1>..:I\\'l'l'n SiZl':-; i:-. :d.·,o impo l1:lIlt l)l'cHI"'c :t fcw I.lrgl· h:lnk, hold ;Iilllo.'t k.tf till' to(;ti ('otnllll'I"(.:i:t1 h:lI,k :.....~l"I.,. and 1"t"lIh", of thl' brgl' hank., m,IY 1ll;1:-;1.. ,how of Ihl' ." nallcr h:lnk~ 20 2.3 7.0 1.2 1.0 1.3 -.3 4.66 3.59 1.24 38.2 1.8 5.' , .7 .8 .1 4.61 3.88 1.57 1.81 •., 2.0 1.2 .7 .8 - .2 5.74 4.78 1.83 74.9 31.5 1.1 ., SA .5 •• .3 3.82 3.29 1.38 2.20 1.10 8.34 .10 7.59 - .22 8." -3.30 .41 6.15 2.23 1.32 3.18 3.42 3.40 66 70 73 78 1.54 8.33 - .18 -2.n I.:!r,gl· hank, :lrl'll1o:-'l' \\ilh tOlal assel:-; l'xn'l"ding;:,1 hillion: mid ~izl" b:mk:-- ha\'l' ;Issels Ix:l\\l'l'n StoO million and SI hillion. and th .... sm:tll·hank clll'gory consi:-;I:-. of hanks wilh assclS umIL-r SIO() milli~Hl Fi,glll"l'';:\ :tnd ·, .~ ho\\· llll· prop()rti(H' (If h:lnk..; :Ind ; 1 .':-'I'I.~ in l':lt."Il .,i".L" cbss. :md Fi).ll.l!"l· :; lkpitb tIll' :-.harp rl'dllClion in nonpnformin,l.! loan:-. for l':lell Si/.l' Cltl'got)' of h:mk T;.hlL,:-.:t I. :Ind :; proddl' (il-I:lilt'd infonnalion aholll t1K' condi tion and IX'rform:uKt' of Ek\'\.'tllh DI"'lrit."t h:tnk~ hy ",izc Small banks, TIK' h:1I1k., \\ill1 k':-;" than $100 mil· lion in a ....~~'I.~. a~ a \\llolL'. 1l:1\'c h ...:l·n rt.·("OH'ring :-.in...-...' l';lrl~' 191'19. Th . .·)' fin:.I I) rl"tul"nl'd 10 pmfib in 1(})(1 1 111..'11' -'1ll;11I 10....' of $:\ million in II)H() IX'(:IIllC ;l profit of51-:\ million in 199(1. In fact. morl" Ih;tn t ilrt.'l'-QU:1I1l'r.-,01 thl' ~!llall hank~ l';tnK'd :1 pmfil Figure 3 Figure 4 Eleventh District Banking Structure by Size Class, 1990 Eleven1h District Banking Structure by Size Class, 1990 (Number of banks) ........ (Assets in b!llioros 01 doIars) (16) SOUACE Of PRIMARY DATA: CoosoIldated Reports 01 Condition and Income (Board 01 Governors, F&daral R&SerVe S)'5tem). SOURCE Of PRIMARY OATA : Consolidated R&pOfIS of Condition and Income (Boaro 01 Governors, for 1990. profits resulted, in part, from a healthier economy, but also from lower nonperfonning assets a nd an improved sp!'<..-ad between the interest earned on assets a nd the cost o f deposits. Small banks' 1m performance, as measured by an ROA of 0.40, is up from the slightly negative ROA of -0 01 percent reponed in 1989. ThroughoUi the 1980s, this group of small banks rema im."<.! relatively stronger tha n the other two groups, but the poor economy did not leave small b,mks unscathed. Still , they mai ntained hea lthy cdpital r;:lt ios ,lIld hea lthy interest margins that allowed fo r a quick recove ry of return on assets. The strong capil;!1 base that small banks reported en;lblcd most of them to endu re the economic downturn. Mid-size banks, 'nle recovery of the mid-size banks was similar to the !'<.."'Covery of small banks. Of the 237 mid-size banks opemting at the end of 1990, 205 hanks, or 87 percent, earned a profit for the year, lip from 70 percent in 1989. TIlese 237 banks produced a composite net profit of $172 million in 19'Xl As with the small banks, this improvement Cdn be anrihuted to an improved economy, a (k:crease in nonperfonning loans, and a more favomble interest sprt.:ad. This group of banks reponed an ROA of 0.35 percent for 1990, an improvement from -0.35 percent reponed in 1989. The equity cdpital ("""dtios fo r these banks also improved, rising 10 6.43 percent in 1990 from 5.78 percent in 1989, as equity capital increased by $300 million. About half this amount is as a result of new capital stock issuance and holding company transact ions; anothe r one-sixth stemmed from consolidations of Imlnch ba nks or OIher business e nte rprises. Large banks, This category of banks with more than $1 billion in assets comprises 16 regional banking giants. These lY.mks also returned to profitability in 1990, after losing $3 billion from 1987 through 1989. The FDIC and, to a lesser extent, the banks themselves, addressed loans that were reported as problems, and this attention led to considerable declines in total no nperfonning loans. The resulting improved credit quality led to lower expenses, panicularly in loan loss reserves. These factors, coupled with the cdpital inj<.."'CI.ions associated with the takt.:over of several large fai led F..conomh: Kevkw - July 1991 FQfal Reserv. System). " Table 3 Measures of Financial Condition and Performance Sanks with Assets Less than $100 Million Other Dislricts' Banks Small Eleventh District Banks Year 1987 1988 198' 1990 1990 Number 01 Banks 1,5n 1.364 1.196 1,052 8.186 SO.7 45.6 20.5 41 .7 17.6 317.3 168.1 3.3 .5 3.2 .4 •• •• .2 29.1 .7 .7 1.2 3.6 .6 .6 • .6 -.2 .5 0 .2 .2 1.3 2.' 4.67 5.02 4.28 4.25 '.00 3.88 3.22 3.02 2.24 2.06 1.93 1.20 1.94 1.53 .80 2.38 8.27 2.09 8.37 1.18 8.52 9.97 .76 8.25 4.07 (In billIOns o( dollars) Tolal Assels TOlal Loans TOlal Nonperforming loans Tolal Equity Capilal Tolal Reserves Total Nel Charge-offs Total Provisions for Loan losses Total Nellncome (Percent ) Nonpertorming loan Ralio Troubled Assel Ratio Net Charge ·off Ralio loan Loss PrOVisions 10 Total loans Primary Capital Ralio Return on Average Assels Return on Average Equity Net Interesl Matgln P8fC8I'ltaQ8 of ProtMbIe BarQ Number 01 Failed Banks .... 29.' 1.4 '3 - .3 2.60 8.53 -.48 -6.35 3.62 24.1 l1;u11-. ..... in(''I'(':.I'I.;(l l hl..' brgl' han"'" capiLli r:1I10.... If Ihl"l' hank, (,':111 m:unl;tin llwir profilahllilr. Ihl..'ir (';'pil;,1 r:llio..... .,holiid oml Hllll' to dUllh ttl llll' 11..'\.:1 of largl..' hank .... I..'lsl·wlll'n.: in Ih...· I.__ Ollntrr thai did not ...·ndllrl.· rq.(!on:11 n:c...·~~i()n~ \'\!hill..' h;1I1k~ in ;lIIII11'l·...· catq(ofl"" " .,111)\\1.:(\ prori t ~ in ]1)1)(1. tl1\.· I:lrge h;lnk~' l'l·(,·O\·...Tr \\:1..... 11K' !llo . . l prol1ouncl..'d. :1.. . b d ...·I11()ll'.t!~lll..'d in Figllr...· 6. Th" group uwr...·:ls<.·d HOA ffOm -0 H IX'ln'llt in 19H1) to o. 19 pl..'l"(,'(,'11[ in II)I)(), Th<.· lX'rc(,·nl:Ij..l<.· of " 3.3 2.' 1.0 .7. - .33 ~.01 . -4.51 -.09 3." 3.77 .'0 5.32 3.69 62 69 77 88 80 .7 107 93 45 3 6 • • (Percent) Failure Rate ., p rofil :!h!...· I.lrg<.· lunks 1'0:-'1..' from :H IlI..TI.·l..'nt in 19KH 10 HI pl.:rc...·nt in ]\)lXI. Ih ...· ill)..tl!...·'1 improH,'Ill<.'nt :lIl1ong Ill...· lhrn' group, (Fi,!.!;//H' 7 1 l>llrillg Illl' !'l·(,'...·':-.IOIl. Illl' brgl' h:tnk'" (,':lfnHlg:-. bad r;t l ll..'n Ittl' funltl·,I. 1<.':1\ in~ room for brj..ll· '" mg.~ ill I(OA A morl' lilorough eX;lllll1l:tllOn or 111...• br~l' h:lI1b IIllltl' Ek\(,'nth Di,. . lrict indl(,·;tI ...·:-.II1:H Ih(" dr;LI11;Llil' (knl';I~~' in nlmpl..'rt{)1111ing a~.""l·IS r...·.... uh ...·d priln:lnl~ fnHll till' F[)IC, rl"ohlli()Jl ()f till' l.trj..ll..' Irollhkd hank>- Tot:li nOlllx'lforrnil)g fcdcn l RC:S<' rvc Hank or D.. u ..... Table 4 Measures of Financial Condition and Performance Banks with Assets Between $100 Million and $1 Billion "",," DIStricts' Saoks Mid-Size Eleventh District BankS Year '987 '988 '98' '990 '990 Number of Banks 307 244 227 237 2.467 69.4 53.3 25.7 (/n billions of do/lars) Total Assets Total Loans Total Nonperforming Loans Total Equity Capital Total Reserves Total Net Charge-otis Total ProvIsions lor Loan Losses Total Net Income (Percent ) Nonpertormlng loan RallO Troubled Assel Ratio Nel Charge-off RatiO Loan loss Provisions to Total Loans Primary CaPItal RallO Return on Average Assets Return on Average Equity Net Interest Margin 34.5 2.2 49.6 21.8 1., 1.6 2.6 .9 8 .8 2.' .7 .5 - 5 .9 -.6 5 -.2 6.50 6.04 4.97 5.12 2.53 4.94 4.13 40 ,., 2.20 2.76 2.85 7.26 6.54 - .73 - 12.38 2.79 -17.44 2.69 -95 1.1 1.90 49.8 21 .2 601 .0 3678 •• 46.9 7.' 3.2 .7 3 .3 3.' .2 4.' 3.9t 3.45 1.45 2.14 6.' 2.9 '86 .79 2,23 7.14 -.35 --6.53 3 .25 1.29 7.65 .35 5.68 3.31 10 .59 424 '06 8.72 .82 59 62 70 87 92 Number of Failed Banks 7 29 34 17 (Percent) Failure Rale 2 , " 14 5 Percentage of Profitable BarKs 11):llb " .... ,L' n:d l1 l ...·d Il~ 11UI1"\.· [11:111 11:111 Irllm 111L' ...·nd 0 1 P)h't) 10 I h...· ...·Ild of 1'.)<)0 "1"\\0 h:lI1l.., in p: rrtillIL, r. Itlill.. OIl\.' "l"l·,.I' :rlld T r.A \1 Ibnl.;. :1t"UIl1Il1 lor :1 1:lr~l' ~h.lr...· o[ 1111' illlprO\ ...·llll·m . Both hanl.;, L·Illl"f~l·d .I((l·r IIlL' I.filml· tll oIIll'r L,rg...· 11;1111.., 111.11 Ih ...· I D I( I"l.·'oh ...·d Iflrol1!-:1l tl1\.· 1"'1Ir('h;1 ......· .lnd .1"tlIIlPlioll 1Ill'lhod - Th ...·....: t\\() bik·d h :11l1.. rl" ( llulltlil' \\l' rl' Irn:l1i/,ed III ItJ'l)(} ..\no(Ill'1' 1T:111'.ldion nll11pil'll't.llIl 19')() Ih.ll (1I11111.: r n.:du ...·L't.1 rl l '\l'lllh 1)''''lrit"l h.I11I.., ' Ilolllx·dimnill)..: Econ o m ic K" "jc'w - J"ly 1'.1") 1 TOIal nonoerfOfl'Y1JfIQ /0,1115 rfochood 10 S I I bllIIOfl from 53 5 I><IIIon Nonperlomwlg rCo11 (!Stale .1 110' fI(It1pe!/oImmg commerc><Illoiins ii'S(] $Il()W('(i ptOt'lOtJtlCf1d 1t11p(0vetnenlS over 1990 OS both fell by moro 1fl.1 <1 51) percent Real e SI<'l'e O.lnet1 ptedclrHtlillCIy /alcclaSe<l real flSla 'e alSO fell no/lCellbly 8.Jn~ On£> re , 1S .. .115 /almt!'d ""m /lit!' M8anJ<s thill ia,1ed r EAM Bani< "',~5 formed Irom tile la led Te oas Arne<lcan "0," Figure 5 Figure 6 Total Nonperforming Loans by Size of Bank, 1987-90 Return on Assets by Size of Bank, 1987-90 6iIIions 01 dollars Per(;Qr11 • , " • , , , -, ------ ----- _..... , SOURCE OF PRIMARY DATA: Consolidated Reports 01 Condrtlon and Income (Board 01 GOVerflOl's. Federal Reserve System) a~-.cts \\';h NCN I-l-l \~x:I:-;' acqu i:-;it io n o f ;l f;l ikxl :-;ubsid i;jry of thL' N:lli()o;1 1 lhnk!'harc:-- Ulf!xlr;uill!l of T<,·xa:. that held mort: Ihan S I billion in ;1S.....:IS '. \,\/ith I\."'tairw.:d l·arninJ?.. from IC)I)() profl's ;1Ilt! fn::--h G.pilai injl'Clion:-.. b~L' ban k .. \\'l·l\.' ahle [0 increase their C'.lpital r.tlio:-. Equil Y capll:ll 10 as."'-·..... inneaM..>d from 2 Ii JX'n.·cnt :II Ih<.' l'nd of 19H9 10 '; 27 perccnt for I')IJO '111L' amount 01 l'(l uity also incrl'ascd by mOl\! Ih:tn 52 2 hill ion: of thi ... ;1111ou nl. 5 I 3 hill ion Glint: from Ill""\' (~ l pi';ll ...tock i ......u\,.'d :Illd !1~Ill:--;Ktion:-. \\ ith p:II\:nl hokh n~ l'ol1lp:mics. i\ suhst;IllIi:l1 portion of this incrC:I."I.:d C lpi!;!) Glllll' from thl' L'(luity inlcction nt :Kk~ wh ..:n Hank OnL--'I'l':>:;IS and TEAM Hank finalizl.!d !hdr l·cSIX'(,'!iVI.· 1;lkeovcrs Firs! 111ler:-.t;l1c l3ancorp also madl.! ;1 cl p ilal il'icClio!l to its .~ lIb... idi:lIl' b;lIlk:;. in Tt:x;!s dllrin~ II)l) J. Room for improvement: a comparison of District and non-District performance Ek\ ..:nth Di~l ril:1 hank:.' IXISi ti "": nl.·' incolllc fur 1'JI.)() (nd icHe.., that 111l'Y :1!"C r..:nl\'t:rln;.: But None bano\s beIc:InQ"'9 10 IIli$ IIOkirng C()'1'lpiit1y were declaredlllSOlvcn18t'ld S{)IdIONCN8- TeX8S()(l.llJne I 1990 SOURCE OF PRIMARY DATA Consol.ctaled Repor!s 01 Condi~on and tncome (6oatd 01 Governors. Federal Reserve SYSTem) Ihl'" n.:t:twl.'l)' cannot h..: lkcl:lrt.'(l l·ompiel": until c.lpit:tl r:tl ;o:-- llll.·L·1 I\!~u latory swnd:rrds and b;mk f;lilllrl,.·:-' :-.uh:--kk·. All Elen.'11Ih 0 ;Slr1(1 hank:;. h:lxc lX'l,.·n oper.lting in ;I TI.'cO\l.'ring l.'conollli(' c!l\'iron1lll.'IlI for thl.· P;I:.I Ihre..: }'L·ar:-., bUI h ank... may nl'<.--d ano l11l.·r full year lx-fore . hl..'ir c:lpilal r:llio::. ar..: nl11:--idL'rt.·d :-.olt nd Thl..' h:lnk:-. Ihal rco:in..d F1)IC :t..:--i."':l ncl.· duri ng their acqui ... iti(111 of fail..:d h:tnk .. h:!\ L' ~c nL' r.tl l }' healt hy relurn .... and Glpit:11 r:ll io... TIll' U ll;l . .... i ..tnl h;tnks. :1." l1oh.:d earlier. h;l\ I.· lX,'uJII)I,.· profitahk' h ut wilh \n;';lk returns. (ii\'(:n tillll' in :1 growi ng economy, however. they :-.hould g:li n financi;ll sl ft:nWh W h il<..· Ekvl.! nth \)isl ri('1 hanks of :tll sizcs showed illlprovl'lllt:nt in 1990. lhe IWO ~roups wilh ;\ .......L·ts k:-.-.... Ih:u1 S I h i llion IliUM improve IUli itef 10 com pare j':I\'ol';lhl y with banks outside till' rL'Won. Thl.! 11)()0 pedorl11;lIlcl,.' of lx)[h :.mall ;tnd m id -si)'.I.' ~roup:-- 1i.·11 Ix'low Ihal for hanks of 1'oimibr 1'o iZl' ollt .. ide thl' E!e\I.·nth Oi... trict Abo. a h ighl'r lX'rCL'ntage of :-.illliiar-:-.i)'ed . olll-o(-J)l.. tri<..·! h:mk .. \\I.·fl· profit:lbk 1tl 19'-)0 T ltl·....... o{hl'r hank:fl·IXI111.'l1 hi~hL'r p rimary l.':lpi.al 1';11 !()1'o; ....mall OUlof-))i .. tri('l h;ln k .. ro'--'Ix)n..:d 9,(r percen t \'cr:o.u:-. K :;;2 ,X'rn'nl 1'01' ,mall Elo.::n~nth [)i ..(riI.1 h:mb, ;lIld l11id-...izl,.' ollt-of-\)i1'oIrin han k!' rqxlttl.'d H.72 JX'r\"L'nl \·l·I'.. U" - 6:;; IX'l'cent for m id-oiizl.· E!t.-\·enlh r..'w F..-dl:l"011 Rc:sc n-'C Rank o f Da Uas Figure 7 th~ir investments frolll lo:ms inlo securities. Total Percentage of Profitable Banks by Size , 1987-90 10;lllS dim inished by nt.'arly $30 hillion since 1987, while ~'(.·lIrities ha\'e inl're:l-.cd hy n~:lfly $ 14 hillion (Figure 8) Severdl cxpl:IO:lIions for this ponfolio shift arc possible Loan clem:lIld hy qU:llified borrowers Ill:!y IX' w(;·ak. On Ihe oti1('f hand, hanks may h:l\,c tightent.x!t heir credit standard... in reslx>nsc 10 the loan losscs tila[ occurred under Ihe tooliix:ral Sland:lrds of pn:violls YC:l rs. Bank regulators. in rc . . ponse 10 the l'OSL~ of the numerous h:lI1k f:lilures of Ihe pa:-.t se\'cr.11 years. may IxCllltioll~ in the cx;ull;n:uion of the loans on each b:mk"s hooks and. in tum. may he requiring banks to he ovcrly prudent in thei r supply of creelit B;Lnks also may he :IIJjusting 10 Ihe new riskhased c Lpilal guidelines hy Illoving into assets, sudl as Rm'crnmcnl securities, thai require less capital. Economic tlieory su).:W::sts that a bank will dloose among (;ornpeti n).: invc.~trnenls based on their (;ompetitivc yields Yields are :llijusted for co.... ts of fund ..... provisions for loan losses. and overhead ~xpcnscs. Umlc.:r Ihl.! simple 6-percenl capital mit'. the risk -:ldjll . . IL'{1 yields for l<xms and .'oCl"uriti~s were :lpproX"imalely \.'<11I:11 With the implementation of risk -hasc..'<.1 C lpil:11 fl-'"qllireTll~nts that c~lI for morc t."qllity carit:l l for riskier assets, SOURCE OF PRIMAR Y DATA : Consohdated Re~s of Condilion and Income (Board 01 Governors, Federal Re5efVe System). District b:mks, II0wl,.'\'er. sj~ns uf fegionaleconomic \\'e:lkness in Ihe other pan .... of the count ry .' illn:tn,--d when ddiflqlll'nI ':n:dit.~ at hank..; in o(hl'f FL-"(Ier:.ll Rc.-.erve DiSl:rich ill(,:n:asl.:o during 1990. Only the rclfonn:mce of the DislricCs brgesl b:mk:-. compared (:1\ or.lhly with thaI of b:mks III other Di:Micts. 13o[h return on asset..; :md n:tllrrl on L-'tluilY were higher in the Elewnrh District Ihan elsewhere: :-.[ill , hoth mea.... ures were below hi:-.torical . . t:ulll:!rd ... for brge banks. Clpil:11 mlios of IargL' EieVL'nlh Distri(·t h:l1lks were st ill below n:Jlion:11 norms. Lmge hanks d . . ewherc :-.taned 10 show we:lkm:ss as their non performing loml ralio and troubled aSSt'ls t~uio increased. becoming ).:rc:Jler than thai for brgl' Eil:VL'lllh Di.~trjct b anks FurL'closed rcal t'st:Ue at these out-of-Dislrict han ks more than douhlL'd during 1990, passin).: Ihe $14 b ill ion mark, Total nonperforming loans for the . . e hanb jumped hy $17.2 billion, :1 37I")!.!rcent incn.::IM:. Figure 8 Total Loans at Eleventh District Banks, 1987-90 BilLions 01 doItars ,ro Lending in the Eleventh District Ik.... pil~ till.:;!' fClllnl to profit:lhility. EI",\"\.'nth Oi . . trict IXlTlb h:l\'t' nol collt:'l1i\'t'ir incrt':!St'd I~ ndin g Oank ..... in fl·.... ponding to pa . . t kXHl lo.... st.'s and n~w rt:').:ubtory rt.'quir~mt.'nts. ha\'e shifted Eco n o mic Kt.'VK:w- Jul y 19') 1 SOURCE OF PRtMARY DATA ConsotJdated AIrports d Contftion and Income (EIoan:l of Governors, Federal Rewrv8 System) " Table 5 Measures of Financial Condition and Performance Banks with Assets Greater than $1 Billion Othe, Oistrn:ts' "ok, Large Eleventh District Banks Year Number of Banks 1987 1988 198' 1990 1990 19 16 14 16 354 73.4 46.1 5.8 27 21 ., 76.4 41 .0 3.0 2.5 1.7 1.3 88.2 41.9 35 2.5 1.5 1.1 893 41 .8 1.1 4.7 1.2 .7 2,268.0 1,424.8 63.9 130.4 43.8 23.9 1.5 - 1.3 1.3 - 1.3 1.4 -.4 .4 .4 25.5 8.5 (In billions 01 dollars) Talai Assets TOlalloans TOlal Nonperlorming Loans Total Equily Capital Total Reserves Total Net Charge-otis Total Provisions lor Loan Losses Total Net Income (Pen;ent) Nonperlorming Loan Ratio Troubled Asset RallO Net Charge·oll Raila Loan Loss ProviSIOns to Total Loans Pnmary CaPItal RallO Return on Average Assets Retum on Average Equity Net Interest Margin Percentage 01 Profitable Banks 12.62 9.85 1.87 7.33 5.50 3.00 830 6.16 2.60 2.71 1.86 1.58 4.48 3.45 1.70 3.24 3.00 5.29 -1 .74 - SO .90 1.88 3.48 4.44 -.44 - 14.57 1.65 1.00 6.48 .4' 12.23 2.52 1.82 7.54 .38 6.75 3.37 81 79 6.34 - 1.74 -41 .65 2.02 32 31 36 Number 01 Failed Banks 0 5 3 (Percent) Failure Rate 0 26 19 2 7 ~· u..'ld :I I\L'I" l'"\lwn"l'" I'" ~rl';lll'r for ."v~lIriliv" Ihan for loan" 'Ilwrdlwl·. h;ml-.:" :11"1.: ;Ippropn.ud) fC"r)(llldlllg 10 till" IIK\'nll\ l' 10 huld mOR' """dld- m.II.. ..· I,.all'" H,IIlI.. 1,,':~ ul. II'''''' uI"gl" h.llll-.:~ ",ilh C l piLll:l1 tor Ix·ltl\\ Il·qllirvtlIlHlHmllm ... ((1 impro\'"'' till'Ll l"qu il} l';IPII;d-IO-,I"'"",, r.llil' '\I,ltl~ h;1I1k... d o Ill'" ;lIld !c\\l'r IO.II\-. ...(. 1)~ .... It '" In~ III" \.\ t.:n n,·\ ~ ·..... ing 111;ltl ).:rt )\\ I h 11ll" Otlll'r tV"l·;lr.. lt "11~~l'''I'' tlt.11 ;1 1;1(:1-.: (II L'qUlt~ l.lpit;t l 1"'.1 n)t\-'l r..lInin~ Llllor in;1 11;111"" ;,h,li" [0 inl"l"L·.'''l· II-. In.111 \ nlunw Cbir .Jlld .., \';'1'- 119911 "ho\\ th,ll. ,lIHong h.ml.." In poor III1.II1l'ial hl";Ihh. Ilw nlill" -Io-.I""'"·!,,, cHio I" ,I "'lgni fiLII1I I.ll\Ol" III hllllltng till' h,'IlI.." ahilil~ 10 \l1l1l1l).: lin ;lI1l"i;dl~ .... '.und 1 .. '"\,1' h;lol.. ... hO\\L'\L'r, C I.lil" .Ind ) l ·.It' IIHllld lh ..· L'qllll\ 1()·.I .... ~·t,., f.Hio I~ tlill .1 "Iglliitlalli 1;lltor itll·"\pbullll.1.! lo.m gro\\lh. ,.,lIg.;.tl· ...IIII.I! Ih.1I ,I' h ;lnl.. ... IX'LlllIl\' k"" 1 .lpII;I I(otl .... lr;llIwd 11ll'~ \\illlll..· itl.1 h,,'llt'l" pO"Ilion [0 L·"\!l'lltl LI\·dll II) 111111h\ l "I(uftl\\l'r... If 111,,-' l.:ha l1}.tiI1).: :.t rurt un..' of till.: banking industry in I hl.' I-:k\'''-'11 tll D istrict con,'II~l ill' Ihl." f\..'gi(lIl',' 1(1:l n slippl y, n)lllI11I."rd:lll'nll'rpri";l's will h:l\'c 10 (k ;1I \\illl Ih ..: ('tlll,'"'qu"-'IKes Blisinl.'s,--.,:s nm,,1 :lti:tPI 10 thl.' Ill' W l"TI.'d il di m:lle by lapping ollll'l" souro,'!"> o f f11l1 d!">, :.ud l a:. h:tn k!"> in o lhl'r regil l lb, cOIllIlK'rl.·ial p:l per, 1)1" 1.'\ I.'n suppl iers willing 10 p rov ide Ir:ldl' lTl.'dil Eadl of 11K's..,: ahl'n1al i\'l's p .....·'CIlIS dl~t\\ h:ll.·ks Borrowi ng from l !tl.' 10C:ll loan p nll.tlK·tioll ofnccs of out-of-SI:1l1." or fort:i).tll h :tnk., hl':ldqll :ll1nc d in olher rl').tiollS I., lI su all y ;t\'ailahk o nl y 10 1:l r).:l· corpO!: Hiolls wilh I( )an dC!l1aml in l',\(t'l '.";S 1)1' 5 1() r11i llil)n Sil11il:trly. Ih ..: c()!l1!l1l'rt.'i:ll p:I1)o.:r 11l:11 kl'l b avail:thlL' 10 on ly thl.' br).(I.:,"1 corpol: lliOIlS 'I'll . ,: 1I.";C of Ir:llk' ncdit i." lI ,.. u:llly rtlOT"l' l'xpcnsivt.· tll:ln h:lllk fin:mcing, In fac!. till': cos\" assol .'i:ll(:d II illl c:ldl Iypl' of :tl lnn:ltl' fin:lncin}.t can Ix: gr"-':lll'" lil:m 1110.""-' :1., ..()(.'iatL·d \I'itll loans from lu(':11 (,.'ol11ll1ncbl hank..;, Ti).:llI l.:lll.:d ('n..'tl il ('lmditillllS arlO nlll iSlliall'd 10 t hc Ek'll'rlth Di!">t rin TIK' n:l ti{ )flal CC<)I1I)111Y has hl'l'n in a !">h:lllo\\ r,,-'cl.'...;siol1 since bl<.: 1990. : tnd h:.mk!"> 11:1liol1\\"id(.' ha\·I.' n..':.pondl.'d wi th a n..'duc,,-'(I. hUI slill po:.llh'<.: gtu\\"lh in loall loll1llK' '" In Ihl.: N0I1hc:I!">t. III )\\I.·\I.T. thl.·I.:I.'(lIl()lnie :.1(IWd()\\"n JI:I." IX'l'n morl' pr0I10Ul1l'''-'(i. :.in illar 10 Ihe SOluhwcst"!"> dOllnlurrl in Ihc 1900s Banks in Ih,,-' Firsl F,,-'(Il.'r:11 R"-·, .... Iyl' U:11l k Di:.lrit."1 1t;I\1..' rl.'(·ortk·(\ <k-clining 10:111 \ 'oluml.' for lit.... p:t!">1 two p'::lrs Thc EI\:\,,,-·nth District '.' Olu look. 1I00\'c\'l.:r. :lppt.'ar," :1 h il hright"-'r Gil'l.'!l Ihl.' Dislrict·:. ).tl·n .... ml "-'l"onOI1111.: pt.:rform:Ult"L', nl'\\' . profil:lhlc kndin).( opportullili,,-'s ."hm tld arise for Ihc ClIIllIIK'rci:t11);lnk." IkcllIS,,-' tllL'!">C I};tllks 1I:I\"l' hel.·n 1'<.:1.'01<..'ring fo r :1I k:.!:-I a full }'l':'lr. th"-')' shou ld hc ahk to fund nl.·II' 10:1n...;, and k'nding :tctil'il Y ... hould i(WrL':'lsl.', )'l·ar... thallk." to thl.' qll:tlity of Ihcir 111:tn:l/-te111cnl. pk-ntiflll capit:!1 l<,:\'d~. :lnd slltlkit'nl <.Ii\'l:r." ifil.:llion, \Vith :tl11pl\: fl.'S,,-'rves 10 adtlre..;s I hc ir 1lt1llperlill'1ningl(lalls. Ihe:.e b:l nks 111:l inla inl'd profi l :lhility. or:1I le:l:.1 mi nimized till'lr Io.,~s, Aftl.·r four ye;l r, :l nd numt.:ro Lis b:mk f;tilun:s, Ihl' 1':1l'1'l'l1th [)i ...;t ri(1·' hank ing S,,-'{10r ft'lLImctl to profi tahilit y. While FDIC :'h" i!">tancl' ~l1 ppOf1 ed the r"-·("(Ivery. thl.' impnll'l.'(1 CCllllOllli,,-' d im;!te tn:lde Ihe lurnaround po:.,~ihk Even ,,,>(). two years d:lpscd hcfore Ekvcnlll Di!">lrlt'1 h:mks lett Ihe dl~·Ll .. of the rcgion',"; rccoI'''-'I)' And I hrce ycars p :lss...-d Ix'li)re harlks return cd 10 profit:lhi lity. Perhaps four or rHore )'''-'; u 'S :U'C ncc ded bl."fofl· h:mks fl.·turn to fu ll l1 nan('i;t! hcalth, Sli ll , thc h:ll1king indll.,lry·s rcturn HI profil . ahilil}' in 11)90 s,,-·nds positivc signals 111:11 Elevl.:nth D i~trict hanks :11\: ft.:llIrnin).: to financi:l l he;llIh :IS wdl. M:l ny b;lnks slill nc,,-'<.I 10 increase their c:lpital !L-\'ch Ix:fore thc fCl.'OVl.'ry (~ ln Ix: dccbr<.:d ("0111pll.:tc Should ddinqucncics t.1Jl1tinll<': 10 dcdin<.: ;md nl'1 int.11l11l' rC111:1in 1")(~,iti\'I.:, th .... EI .... \"cnlh D istrict's hank." (~1I1 proc<.:cd down tllo..: road 10 full r"-'('()\'t'IY froill till' l unooil of Iho..: 19BOs, Conclusio n 11\ I')H(J. 111\: Soul !l\\('sl',"; ,,-'Conomic do\vnlun1 rcslllt ...·d ill dr:l "licl l ly i nt.'l'l·:L";I.'l1 nonp.,: rfo rming loa n:' :l1 :.011 1"-' Ekl'''-'nlll Dblricl h:lI1k~ SOll1l' hank m :tnagl.·l"!"> \\·",·rl.· 1I 1l:.1hlt.- to fOrl.'ClSI :'1t.'n l1~u...-1)' thL· sl'\'crily of tlK" downturn o r to lK"dgl.' adl'qu:lIl.·ly :I).t:li rbt il!"> dli:l.·t, j\ l :Ul~' ., ud l in:.tillllions l"vqllirl.:d fl'dl'I:11 lIl!<.:ry,,-'ntioll w lwn Ihdr cllaf1crin).: :llllhorily {kdtfl.'(I,IK'1ll ill:.oh-cnt. Olh<..'r in ..,ilUlioll!">, hOI\c\I.·r. wert: ;Ihk to Cl JIltinll"-' ()Pl'I:ll i()1\"; thrtlll,gh{)ul Ihe lurhllknl Economic ikvll'W - July 199 1 TOIai loans ar our<JI·{)rSlfICI DalIllS actually declined ONer /116 fourtn Quat/Ot 01 1990 21 How Banks Cope with Nonperforming Loans At every bank, the management expects and prepares for nonpertorming loans. In any economic climate, some businesses and individuals to whom a bank lends will be unable to repay their debt in full. Bank management attempts to minimize loan losses through its underwriting standards and periodic loan review. Bank management can take several actions once a loan becomes delinquent. The bank could attempt loan collection, incurring such costs as legal expenses associated with foreclosure. If this effort is unsuccessful, management could write down the value of the loan on its book to a level that more accurately reflects the loan's collectible value. Management prepares for the eventual loan loss by establishing a general loan loss reserve account to absorb future losses on loans. The sizeofthisaccountdependsonmanagement's expectations of future loan losses. The general loan loss reserve account, called allowance for loan and lease losses, is built up over time by retaining income. Additions to the loan loss reserve account, referred to as provisions for loan and lease losses, are noninterest expenses that reduce current-period net income. All loans are priced to cover all the anticipated costs of making the loans. These anticipated expenses consist of interest costs of funding the loan and a share of overhead, which includes loan collection costs and provisions for loan losses. As long as the amount of nonperforming loans remain at or below expected levels, bank management can handle the delinquencies internally without a threat to the bank's capital position. If nonperforming loans are greater than anticipated, the bank must make higher-thanexpected provisions for loan losses that cut into net income. Because these loans are not generating revenue, making the necessary additions to the loan loss reserve account becomes difficult. These loans also generate other expenses, such as monitoring costs, not needed with performing loans. Continued losses from nonperforming loans will jeopardize the solvency and viability of the bank. The reserve account that was counted toward the bank's primary capital requirement becomes depleted.' Because banks must maintain minimum capital-to-assets ratios in accordance with regulatory guidelines, banks that are at or below these minimums typically respond to a drop in primary caprtal by raising more equity capital, making additions to loan loss reserves from current-period net income, or reducing assets to the extent that the bank is able. When these efforts to recapitalize fall short, the bank is insolvent and faces resolution by the FDIC. , Before December 31 , 1990, regulators counled loan loss reserves as an element in pnlllClry capllat A5 of that date. commercial banks are rEtqu,red 10 meet capital requirements that are based on the risk leyel of their assets. ReseNGS count toward supplemental capital under the new rules (AHriend) References Alfriend, !l.lalcolm C. (1988). ~ I nternational RiskBased Capital Stancbrd: History and Explanation. ~ ft.'C.Ier:11 Reservc Bank of Richmond, Economic Remelt'. December. 28-34. Comptroller of the Currency (1988), &mk rai/ure: All Eva/llafio" of the Factors COlltributing to the Fai/ure ofN(lliOfla/ Banks (Washington, D.C.: Office of the Comptroller of the Cur- rency, june). Benston, Gc..:orge J.. Rohert A Eisenbeis, Paul M Ilorvitz, Edward J. Kane , and George G. Kaufman ( 1986), Pers/X'Clives 011 Stife ami SOl/lid Ballking ' Pasl, Present and FI/lure, (Cambridge, Mass.: MIT Press.) Clair, Hobert T. (1990), - Performance of Eleventh DislriCI l3,mks in 1989: Progres.,> but Not Profits," Federal Ikscrvc Bank of i),llbs Ecollomic Review, Septemher, 15- 24. Frtlerdl Deposit Insu .....lnce Corpor::llion, AtHmal Report, \';Jriuus is,,>ues. Robinson, Kenneth j. (1990), "The Pcrfonnance of Elevent h District Financial Institutions in the 1980s: A Broader Pe rspecti ve," Federal Reserve B:mk of Dallas, Finane/a/Industry Studies, M;ly, 13-24. Short, Gt!nie D., and j effery W. Gunther (1988), - - - , and Kevin). Yeats (991), "Bank Capital and Its Relationship to the Credit Shortage in Texas: (Paper prescnu.."(1to the Feder.ll Rescrve System Commiltt.'C on Fi nancial SinK" ture and Heguiation, San Antonio, Texas. january 8). Coldwell B..'lnker Commercial (l5)CX)), Co/du¥">f/ i3a1lJ...'erCommercia/ Office Vacancy Index of the United States, December 31. Economh:: Rl:vicw - July t99 1 7be Texas 7brifl SillUllion' /mplieationsfor the Texas Finane/a/Industry, Federa l l{eservc B;lnk of Dallas, September. Y(kd, Mine K., and Paula K. Tlicker (991), ~Thc Rise lind Fall of Oil Prices," The Southwest Economy, Feder:tl Reserve Bank o f Dallas, March/April, 1-3.
@FedFRASER