Full text of Review (Federal Reserve Bank of Dallas) : July 1990

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

FEDERAL RESERVE BANK OF DALLAS Ju~  •  1990  •  cononuc eVlew Identifying Service-Sector Exports from Major Texas Cities Robert W. Gilmer  The Texas Index ofLeading Economic Indicators: ARevision and Further Evaluation Keilh A. Phillips  This publication was digitized and made available by the Federal Reserve Bank of Dallas' Historical Library (FedHistory@dal.frb.org)  Economic Review Federal ResBrve Bank of Daf/as  ..  IkIbel1 H. Boykin -~ OoeIf.of<IIlI>IOOI,...  William It. WlIlI,ca f.,r lfao " ' - - WId a..I ~tI..,. 011_  Harvey Ros.nbhm ~  vahfidMI atd  ~lIIf1nntr1.  "'--  Gerald P. O'Driscoll, Jr. "'--!JIo"«IG<"'~b  W. Mlchul CO)!  ............  Vi<O'' ,._1M  Stephan P. A. Bnwn .u..s-1'Iar~MId  -,.,.."  Econoalists NaIIOfla! and Inlernarional John!: Hili Evan F. Koemg Hoben T Ctalr tara S Lown  Keonelh M EITI!IfY Joseph H Haslag linda C Hunter MafkA Wynne  Regwnalaro Enetgy William C Gruben Roben W Gilmer  Mille K YOCel Keith R Phillips l()l'i l TaylOf FiOfla 0 Sigalia  EclitOfS Rhooda Harris Diana W Palmer Virglnra M I\ogefs  The fcwotmt; RfMew IS pubhshed br the federal Re_ Bank of Dallas The vrews e~· pressed are (!lOSe 01 the IlUtIKn and dol rIOt nKllssaflly re~ecl the posohons of the Feoeral f\eseM.! Bant 01 Dallas 0' the federal Reserve ~,~  SUb$tIOllIIIJIlS are ilY301ab/e I,ee of cl>¥ge Please scod reqo.oesls /0'" s,~le-copy aoo ... ,1· lIpie-ropy lUbsCIIPlu)OS. bad< , _ . and ad·  dress changes 10 the PIIbI" AffBl1l OepilfUnoot F~11Ieserve  I/Is,  Baok 01 Dallas, StatIon K. Dal-  rel\.ijs7S12l.(2141651~  ArtICleS may be rllpo'ioted on the condill00 lhi1l lIIe $DUI'ce 1$ credited anCIlhe Resean:tI [le.  partment IS pn:r.<(ied WIth a copy of too IIIAIIlta· tlOO CO'llau"OOg the reprlll1ed III!Ilerial  Contents Page 1  Identifying Service-Sector Exports from Major Texas Cities Robert W. Gilmer  Hobert W. Gilmer uses" new analytical tc<:hn iquc to explain Khr longtime ri\·;II.~ Dallas and Houston can coexist less than 250 Illiks apart. This nl;:\\" lCchniqul' permit" an unusually com plett' analysis of the roll' of St;n";ccs in Lhe growth of dties. TJlI.: analr..;;s hx:u...es o n distinguishing  scr.iccs thai gent'r.JIt: ~f()\\ th as exporL"; to national ,lOci intern:ll ional markl'l~ from St'I"\";CCS that arc ll.~d by tht' cit)' and its surrounding ;lre:1. A.o; a ca.~ sludy, h(,' analyze ... S(.;'IY;Cc-se<:lor exports (rom IlouslOn. Dalkls, Fort Worth. and San AntoniO. Gilmer'S :lpp1"O::1ch depict:> 0:111:1." ;mtl Ilow,lon  :I:.  complementing,  "nher than l'()lllpetin~ wilh. c;.L(:h other. D:III.I~ b the distribll~ tion and firumdal centcr of the Southwest. while Houston is tilt.' n:ltion'~ prt't:'IllIOCnl oil <,:entcr.  Page 17  The Texas Index ofLeading Economic Indicators: ARevision and Further Evaluation Keith R. Phillips  In this anide. Kdth It Phillips ro::visl's lhe Texas index of ll';lding eCOIlOIl1i{: indic.::llors thaI he introduced \wo year.... ago He docs so in respollse to recent structu ral chlmgcs in the state economy and the iI\ "iklhililY new data. The  or  Fcdeml Neser\'!: Hank of Dallas h.l~ prodllc...'tlthe Texas leading index Illonlhly since JlIly 19H8. t:sing a m:wly d~•.'n:loped technique for e\'aluating leading indcxes. Phillips find:. lhal the reviscd Texas leading index has Ix·,fonnt'<.! \\'dl in predicting movClllents in [he T...·x:ls t'conomy ~inn.' 19R1. He also IIn<.b; Ihat momhly n.'\-isions in the leading mdex arc gene"llly ... mall and lhal pn.'liminary eslim:lles arc Rood predictol'S of IInal "alues" TOAl·lht'r. these re ... ult.. indic;tie Ihal Ihl.' new Texas leading index can 1)(.' a u~flll tool in impn)\"ing fo rec.::lsls of the state":. dynamic l,{'onom)" "  Robert W, Gilmer Senior Economist Federal Reserve Bank 01 Dallas Houston Branch  Identifying Service-Sector Exports from Major Texas Cities  A  s ClllploYIllt.:l1t i n large cities stcad ily shifts from manufacturing 10 the Sl'lyiCl' sector. a ljlll:stion th,n hecomes inn easinl-(I} imponanl is wheth(::r till ' selyice sector initiates growt h or passi\'cJy rt:<lcts 10 exp:Illsion in othcr Sl'l'tor.... [n many met ropolitan areas. cspcciaJly among tilt.: brge,"t ones. st.:f',ices do initiale growl h. [n Ollt: of tilt.: bt.:Sl known hooks on the subjr.:cl. St.l nh.1CI.; and r\oydlc ( 19H2) nlllcJ ude -tilat mClfopo!itan r.:cono m it.:,'; ilfC signifiGl ntl), spedll lized .. .(.md l th;11 the term ... of Iheir sr c<:i:l lization arlO inlTeasinglr d d ined hy d iffere nt mixes o f scl.... icc (as oppost.:d 10 nonselyice) lICl i\'iIY," Thl' purpose of this artide i... to exami ne sClyiLc'Sl'CtOl' expoll s fmm fmlr T exas metropolitan area." with a popu btion of o ne million o r morr.:: IlouslOn. Dall;ls, Fort \,'011 11 . ;lntl San Antonio. 1 Exports art.: irnponant heGIUsl' a region's export s arlO !inkt.:d to its :Ihi lil )' to iniliate WO\,·th . If St~H, hack and Noyelie are corrcct. thb lisl o f exp<llls-sectors tilal dcllnl' the tt·rrn!'. of IOCII scI'\'ict.:,sr.:ct()f spec i a l i7.ation-proddr.:.~ illlportanl i nsi~hts inlo tht:s(:" (:tt ic!'.' economic lift.: :lnd their rolt" in Iht" American system of cities. I use ..I newly dt:\'eloped ml'thodol()~)' (0 identify k e)' sely ice st'Ctors within an urhan SCI" \'ic(:" econolllY , Ilat gt:f)er;.Hl' exports .mtl impoll s. Selyices net:d not bl' t"x[>011ed al aiL of course, as c\'ery city i." lil.;dy to h,,"e its compkment of purely rt;:,<;idt:lltbl selYices. such ~IS bundry or shoe rep air. Even when sen'ices <lH,' <:xpol1t'd from a city or metl'Olx)lilan 'Irea. the}' can hI.' di, \'ided into 1\\'0 pailS: those exroll.~ sold to an adjan:nt h interland of sm;llier Mlrn)lInding ci(io.;.<; ;lIlel nt:arhy to\\'ns :lIld adjacent rur;1 1 'lreas. and (host" eXfXll1s st:n t more Widely to other ci ties and I-: cooomic Roevlew - July 1990  mt'lrorolitan arcas lhroug ho ut the nation. EXPOl1S into SlllTOllllding hinterlands :m~ ;1 (-ammon f<:<I ' ture of all hlrge cilie ~ . wi th expol1s of n:..·w il sales da large urhan shopping celltl·I'!-. ;1 com lllon ex' ;Impk. Il o\\'e\'(~r. st.:lyices sold w iddy throughout the st;lte Dr nation to other largl' ci ties ind icate urhan c( )m raral i\'l' ad\·;tnt,lge and srecia! izll tion in Sl'lyiccs, which is ~'hat I seck to sh ld y in th is article. My method ology allows me 10 identify Ihese widely tl:l(kd selyi(.'es . T ht.: resu lts o f th is analrsis pl"()\'idc a broad pel'sf>!.,(·t " 'e o n l hl' imporl<1nce o f selyice industrie.<; in the economic life o f c ach Texas city. Oallas, for example. emerges as a hi~hly dh'crsified regional dbtrihutio n centt:"r pro"idin~ whoksale, rt'wil, ;lntl Ir:.tnsportal ion SelYlcCS throug houl the SOlll hwcst ;tIl d ti K' nation. Houston. in cOntr.lS!. i., a major oil center. ;lnd en'n ils scl .... k e se('tor appears to I~ domi nnted hy its rolt.: in the oil ;mel ga." industry. Houston's sel.... il'c-.'il·ctor exports ;lfC mostly linked 10 o il :tntl ~;_ s transportat ion. c n~i­ net:rillg sl' n'ice.~, o r cncrgy-fi nance ,In i\'itics, FCII1 WOllh ;md S;1Il Anton io, each \\'itll ; 1 popul;nion of 1.3 million . arc smaller metropolitan areas than either H otl.~t()n (3.1mil!ion) or Dalla!'. (2.5 million). ,lnel ser..-ice-sl'cto f exports art' fo und outside of somt" h igh ly specialized niches. sllch <I!'. military hases in San Anto nio and mil  rc,,'  The term expon m reglONtl economICS refers 10 sales to buyers /(X;Med outSide Ihe reglOf1 undet study not neccs' safl/y Iocafec outside the country rlw bUYe! may well be outs,de lhe country. /)(II IOf purposes oI/h,s paper IleOf she lI'l8y eoually  welt be /ocSled ,n SnoIhe< large melfOf}Ol,(an  area such 8S PhoenIX Of ChICago  l'onnt"ctions in Fort 'X·ol1h. ApP<lrl'n1lr industria l spL'ci:ili;.<.atitm \'ia thl' ....L'I'I'icL' ....L'l'tor. as \\'0:11 :is thl' ahility or .'iL'I.... i<.TS to sh:I)X' thL' L'(,(lnOI11Y of:1 1111::tropolit:m arca. is rapidly lost :1'" \H: l11on' frol11 tht' largt'st metropolil;m are:ls to smalk'J' ont.·s. The first t\\'o M.'(1ion:. of this aI11(')(.· demon, .... tratL' a technique for identifying Sl.'1'\ in·-!'ot.,(.·to!" e:qxlrts. This require:. using MIIllC eSM.:nlia l l'Olln'p!s. slich as thL' eXpOJ1 ha:.l'. loe::llion quotiL·nls. agglomt:r:.llion economiL·.'i. and tilt' lIIiXIll hilTJrl·hr. Tht' third sect ion tLlm.~ spl.·\:ifk:.lJy to till.' expol1s ;mel im pol1s 01':1 Irpi{·allaf!-(.: lllelropoli· tan ;m.:a in thl' l'nill'd ~t ales ;1I1d t'(Hnp:lI.....S it to similar data for Tl'x:l.~ dlil's, An APPt.'lldix pro\'ides Jllctiludoiogicil deWil  The economic base and the location quotient A cit~· Ih'es hy ils CXPOl1s For cx:unpk. I10uston sdb oi l and natur;11 gas Ihroughoul th<.: \\'orld and usc." the:: prtX'L'l'ds HJ par for ils imp(l rt .~. Sl1(:h ,IS :Iulo." frOIll D<.:t ro il. ;Jnd for inh<.:rcmly k)c:11 st'n·ices. sllch :IS dry ('IL-a lli ng or SIlCll' l'I.:p,lir. F.xIX)lts g<.:l1e::ralt' i n('oI1K' frolli ollt... id{' til\.' region. and tllt~y hl't'OI1lt' the kl'~ In long'H:nn growth and dt'n:,lopnl<.:nl. TIll.' indu...lries Ih;ll prodlll'(:' t'xPOlls ;Irt' t.•.III(::<.I (Xlsic St.'(·I( II'S. and ('()I· lectin:l}' ther ;Ire the:: e.\/>OI11X1SI! or Ihl' fl..'gion. Induslrie::.... thai do nOI L'Xlx)r\ hul pro\·idt.· g(J(x1_~ and ... ('rYil·e.~ on ly kK::.tlly art;.' 1J(JIII'JUSi(' IAndn.:\\'s  1953--;-;). As di.\l·us.'ot:d :Jho\·L', Ihe pUfpOSl' of Ih is ;Htidt' i.... to idt'nl if}' h'l.~ic or \:'XpOI1 ....(::clors I hat ari.:;\:' \\'il hin tht' st'ryil·I.' st't'\or of indil'idual Cili<.:s IlowC:'\·er. ht'yond::'1 fl..'\\ ."clttt'I't'd ;Irld O(':c;lsi0I1;11 sul'\·cy .... d.lIa tkll t'xplilitly idt'rltify l'Xporh hy citi!.!s o r other regions of till." l 'nitl.'(l St.lles do not exist IkGIUsc.: of this I;Kk of d;lta . ;Jnd IX'C;JUSl' thl' identilk"IIion of t'XpOil ."t'Clors is nudal to fl" gi()I):l1 c(:onomics. l"conoll1ist .. han..' dl'\'L'lolX'd it nllJlll~r of em p irical ,Ind jmlgll1t'IlI::'11 IL'L'lmitjllcs to Sl'»;Jr;J Ie h;Jsic from nonh;Jsil' Sl"CIOI'S (Leistritz ;lntl .r..lurdo.... k 19H1). The :1j>pro:l('11 i ll this ::.rrlil'i(: i.~ ,11C popular leIClli{)Il·(llIOL .... llt tt'cillliqlll' \\'itll ;1 mo<.!ifk:llIC)!) to isolatl.· 1I1h:JI) I.'xports to ;I[,.'as Ix'yond the city's hintt'ri;ltld n'I.'il and ~l:Jck 19H6). A lo(:ation qllolit.'nI (l.QJ is a simple llll'a."urt,' of spatial ('(mcelliratiOIl.  l  =-,:-'-,"_' ' I '-,-"  LQ  'c'  1/  ]' ~,  r,..  wlll'rl' I' - 10(.';.11 l'arnings in industry I and ," pl:lCl' j .1 :~ 10\;11 e::arnings in pl;Ke j l ~. - rt.'~ionwkll' elrnings in induslry l'l'gioll",idc L'~l rninR~ in all indll .... trit's  -  1:.,-  A loedit)' with an LQ of I i.~:1 Irpic:! 1 pl:JL'l: in thL' ,h;1\ illdllstr~· i ('olllrihllle..·.... thl;' sallll' .... h:!rt.· of c:lrnings t(l lht' locallt~· j ;lS l() the f(.'gic)J1 ,IS a \\ 110k': an LQ \';Jlue of 2 indicate., ,I lex'al ... hart.' of indll .~try l':lrnings I\\'i('(' Ill\.' typical pbn'. :1I1<J :m LQ of O. ~ indicates tlwI this indllsl~ nml rihult':. only half the nonnal sharI.' of c;lrnill~~ to the..' lex·ality. 111(,' Iypic:!!. n..'gioll\\'idt' pl:in' in thl.' denominator i.~ cOlllpriSL'l1 of all brgl' :lIld .. ma ll cilies, town.... \'illagcs. and adj;l\.·ent rtlr:ll ::.m;:as Olll· intert'st in thl' leKatioll q llot iell1 ce::nlel'S on thl' 1':1('1 thaI if t;IStl' and tct'illlolo!!y :I re tht' .~: II11 C in all pl:Jces--dti('s. to\\'n:.. :1I)d \ il bgl'.'ithen {hI.' LQ.~ for nonhasic.' or inh\!I~lltly IOGII :Klidly I.·\\.'r~ ...yhel'l' shou ld (.'qual 1. That b. umlt'r theSt.' drnllll.~lan(e:.. t.>:.lrnin/-!s for nonl);lsic indu...• try (·onlribuLI.' IlIl' S:I111l' sh;lre of carnin~s in I.... ·c~­ rt~ll·l' . S"'·l' till' Ap]X.:ndix for a simplt' pnlof of {his f;lct. In plal'C<; that expor1 a '''l')l..'cilk wxxl or ....'..'r\'i1.'I.'. I \\ oliid t'x lx'('\ 10 fin d ;1 11 I.Q ~ign i fi GlIl l ly ~n:alt'r Ih::.m 1 for 111;11 good o r Sl'rYk't' 1x."":::IlISl' e;Jrning.... anTlK' from hOl h IOGII ~::.des and from e"pOll.~ 10 olllt'r :m,::!s. In plal'''''S Ihat illlPOll W>txts or sl.'f\·in:s. the I.Q i~ less than 1 IX'GllISC import... disp!ae.'L' ]o(:al earnings. :'\oni>;I.... i{· goods :lnd .'>l·r· \'I(·c .... \\'ith I.Qs {If I. :Ire IhL' middlt' ground, which indicates self-sliffil'iL'Ill,\,. To identif~' h:l.~k' :lIld nonha ... i( industrics. I ProIX).'i(.· Ihl.' follo\\ in/-: nlk': Look at thL' Illl';11l ,md sl<lmi;lr<i l'l'I'Or of Ihe leK':l1iun quotit:'nt 1'01' an in· dustry. ml.'asltl'l:'d :I('I'OSS;I 1:11'1-\1:' nllmlx'r of pi:lct's: nonh;l.~i{· imlLlslfio.:s :lfe tI1O'... l· with ;1\ I.·!~lgl· I.Qs IlC'IJ' 1 anti \\'iLh a small stand;mJ dt'\'iation LQs lor hasic inulI ..tries h:u·e iI Illuch i:lrger .... t;mdard ""rI'or. a .~ tht' paltem of (' XpOr1~ :Ind impol1s im, P:U1S ;1 1,ln/-:t' of \'allle~ :lhl.'rn:IIl\·l'ly :lhO\e and Iwlow 1. S(.'I1."{'  Federal Resc l'Yc B:lnk of Dall~  lion Of mort! inhabit:tnts in the CniK'(1 St:n~ .... in 1987, and in principlt' the)' all sit on this lOp tier. The next tier i.... labdt'u ~ ....mall metropolitan areas: and Ix.'low it i~ a tier labelc.."(l -Iowns and "i1lagt'~,~ If thc world were a flal plane. tr.msport<ltion ('O$ts were uniform. ;md pcrf(:ct competition prev:tilc..'d. the pattl!nl of two small metropolitan are:ts and . . ix non metropolitan are:t.' hdo\\' e:Leh major mClfOlxllitan :Lrt':t can fol low fmlll .~t:lndal"d resull~ in the economic theo!'r of how {·cntr.ll pbccs of diflcl"clll .~ize rdate to one :Lnolher (King 1984). The real workl is mort' complex. of COlti"SI.·. and our ao<tlrsis really depends only on !'it! I);lr.lling tht' tor li(:r of htr~e metropolitan an.:;!s from thc rl'st of the hier;m.: lir. Thl' exisl(:nce of:t hier.Hchy means the IOGltion of $ef"\'ict' industries is no longcr a decision 1ll.ICIL- indepemkml)' of c:teh celltr.11 place. Tn particular. m:.tjor llrlXLn areas mar offcr some busines.'><::s external cconomics through the conet'ntr.t· tion o( I..lrger numbers of OIher. related industries in Ihe :Irl'<l 'l1leSt' (,...conornies, oftt'n c<tlled (>(;0110mies (lj"(lI!J!,/ollt(>ralioll. may facililalt: huyer COI1,'cnienee IhnJllgh tht' proximity of shops SI,.'lIing similar !llcl"{·handi . . e; or thl.'y may allow thl.' 11m] lo dr.I\\· from larKc poo ls of SI"X"Ci:tlized lalx)r o r  Urban agglomeration and the location quotient Sollie:' rea...on ... :1 nonhasic location qllotient might as.... ume ,·"llIe.... dim.:r<:nt from 1 han: alrt::ldy I-x:cn su~estcd : taste for p:llticular product... 1lI;1~' differ from one plact:' to another. local tl.:chnology ma~ dillt'!" in production. lind diffeH.:nt income le\"(.~1s :IITt'<..l <..·on ....umption plIttr.:rns. Tht: problem. of com . . e. i.~ that the . . e sollrce .... of , ·:tdation make il ditlkuh 10 :Ippl~ the nit.: we prolx)sed 10 se:-pa r.lte b;lsi{' and n()nh'l.~k industries. "-hen dot:s '·ariano.:' in 1(""lion qllolknL<; result form exporlimrol1 p;lllern~. :uld when do olher fa("(ors GlliSe LQs to nil')"? I will recognize ,tnd cope wilh some of these:- problems on :1 case-hY-Glsc hasi.... in our analpis of Tcx:LS (·ilies. "111is st'clion. h()wt" ·('r. prescnt.~ another important re:t . . on LQs can differ from 011(' [(Kalil)" to anothcr--{hI.' IlItJall bleJ"(lrcb)'<lnd Ihe surround· inK hinterland ."l"I,"('(1 hy l;trge lllciropolit:ln pl;i{·e .... Ch,I11 1 dt.'pkts Ihl..' mil,1Il hier.ln.-hl' that I impo~ on this an:tl)'!',ls. The top tier. htrgt' l11t:'tropoUlan art';t ..... include.... lht' four Texa.~ citks examined to {hi..... u1ide. In f:lCl. this analysis will (:ncomp:t ........ all 4.. l11t!tropolil:m :lI"Cas with ont' miJ-  C hart 1  An Idealized Version of the Urban Hierarchy Large Metropohtan  Aleas  Small Mellopo li l an  Aleas  To~  ,,"  Villages  Economic R(.'vh..-w - July 1990  ,  other inpuLs (Meyer 1977). The result is often a n)llcentr;ltion of many forms of business at tht: (Op of the urhan hierarch)'. resulting in the Sliles of t h e.~t' products by large cities to smaller cities and Ilonmetropolitan areas. Because of these agglomeratin' Lx·onomie.... \;lrger metropolitan afe:lS exhibit IOGllion quotients ];lrgt'r than 1 in these tines of btl.~i n e.~~; ... maller phlCes have smaller LQs. Onu.' mort.' the que_lion ;lrises: When is th is \"ari,Ilion in I.Qs the result of :1AAlorneration and when is it due (0 expmts and impons thai arise from some compar;:uivl' a(kant~lge derived from factors other than heing in a large area? A correction for these t.'conomies of agglomef-;uion is re;ldily ;I\·ailab!e. In eITecl. I drop allihe Slllll ilcr placl's (the sm;l ller cities, towns. and \'ilhlges) from our analy.<;is and ill.lke the basis of comparison for LQs all large metropolitan :neas. The recomputed location quotient will look like  now eliminated. (2) I isolate eXf'Xnl'; and imports by looking for those sectors in an urban area where the new location quotient is considerably larger or smaller than 1. The paHern of trade indicated by these adjusted LQs is strictly among the large metropolitan places across the top l<:."\-·el of Chan I . Exports into the hinterl,md , driven solely hy <lgglomeralion, are now removed. (3) Changing the base of the LQ genermes information about the pattern of s<llt:s fro m large places to smaller places lInd vice versa. These are the sales Ih:1I mO\'e from the top to the bottom liers of Chart 1 (and vice \'ersa) rather than across tht: top lint' of major metropolitan areas. I can tell, fo r example, which industries in OlelrOf'Xllitan i\re;l.~ are strongly subject to agglomeration by looking at the largest shifts in the LQ bast:.  thi.~ :  A typical major U,S, metropolitan area  LQ' =  y,,,/ y,~,  " ();,- )',,)/()~,- y,.), when,' .\'''' - emnings in industry j and a largt' metropolitan place .1;.., .. earnings in all br!-te Illetropolitan placcs .1'" ., earnin!-ts in industry i in ;111 slllall metropolitan or nonlllt'tropolitan places .1:. - t::lrnings in a ll small Illctropolit<ln or nonmcLrof'Xllitan places }~,. );.. - ,IS defined "hon: for alJ places in the region This suhtl':tcl.'> all pl:tces sm;lllcr than a major metropolilao area from the b,lse of the IOGHion quotient. The typical pbce used for comparison in the hase of Ihe LQ is nm', the typical larf{e mClropolil:m place in the United States. As discus.'>ed more rull~' in the Appendix. {his shift in Ihe base of Iht' LQ rt,\,t::lls thr!;'!.:: nt',," ,md illlf'Xlr1ant types of informatIon. ( I) If Ihe new location qUOIients are computed :lnos.~ all largt: lllelropolit;\I1 areas, Ihose that h;tn,;' an ;lI'erage nlue of 1 .md a small standard error are from <I nonbasic induSll)·. The proposed rule to separatl' h;:lsic and nonbasic indllstric.,; is :Igain oPl'I~I{iona!. I still h<lve to put aside differences in taste and technology, but diffefence.~ imposed hy urban agglomeration art'  This section provides gener-l! background and information on data sources and on the level of indllstry dewil available to an:"Ilyze service.~. Also, hcfofe turning to major metropolit,1O areas in Texas, I look at the typical rn:ljor metropolitan area in the Lniled Swtes. For this typic,ll cit}', I focus on the service sector and identification of service industries that frequently opcmte as export sectors in American cities. For purposes of this study, a major mefropolifall area is defint:d as a primary metropolitan st;lti.~tical are;l (PMSA) \\'ith a populmion of more than one million in 1987. Tht: largest PMSAs ,m.: !'\ew York (8.5 million), Los Angdes (8.5 million), and Chicago <6.2 million): the smallest of the PMSAs that manage to exceed one million in population arc Ilanford. Charlotte. and S~ll t I..;:lkc City (all 1.1 million). Among these 44 PMSAs, four are in Texas: Houston (3.2 million). Dallas (2.5 million). San Antonio 0 ,3 million). and Fort Worth (1.3 mill ion). The data for the metropolitan areas arc e:lrning.~ in 1987 by two-digit industry defin ition within the service .~ec(or.z Earnings are a hroad meastll"t' of lotal cOlllpcn.~aLion for \Yage and sal.try employees: this tl\"o . .diKit level of industri'll detail excludes proprietors' income, The senJiee se(:torfor purpo~es of thi.,> paper consiSL~ of: Iran.~­ pOlt<llion, commllnication, and puhlic utilities Federal Reserve Bank of Dallas  (TCPl' l ; whole!-oal(;' lmelt:: rel:lil Ir:.ld<..': finanu,-', in..ur:.lIlCt!, :md rl':11 cM:Ue ( FIRE I: pc.: rsonal :,nd hu!-oines... .'i<..'I'\il'e ...; and Am ernm(;'nl. St't.' Ihe left ('o lumn 01 Tabk 1 for a liM o f t\\'o-di!!it industrie!> in ca~h of thcw <"';l te~orit-:-, T hl' '" C)-d igit Icn..'1 of dt:tail must he h'pi in mind III thl' inlt: rpn: t;llion of result ... For l'xam plc, thi .. ]"" 'cl o f det.. il wi ll allow li!-o to idenlify :1 city a:- h:1\ m~ CXr<)11 .-.t r<"·n~th i n husine:o.... st'r\·ices. I lo"·cn'r. m(Kle... t exports in compu ter s<"'r\'i~es (;1 subsel of bu .. ines,,, seIYi(:c.,,) l1li~ h t Ix: sw;unrx.'d h y dd'ici<..'nt'il':- or l" '(: O ,l\'l'r;tgl' fll' rfofll1:mn:' in ol h('1' h usim':-s ..,ef\,1t"<..' S(·('tor:- Still. :IS thl' n,'sults of I his al1ic.:le indicate, thi:- 1L'\'d of deWilcan p rm ide :1 good picture of;tn area':- :-er\,il'l',sn·tor si rengths anu wc:, kncsM:s. \ Effects of urban agglomeration, r\gglom l'l'llion of soml' hUsinc,"'' 'l'.'' i n major llll.:tnllXllil ;l n ccnters <..·:In (·aU." l' nonha:-i<.. location quotients 10 I>t.., gn.:<tt<..'r than I w hl 'n compult.'d on a h:l ~ of all p b ces in thL' urban hll'mrchy ( LQ1. Computing thc 1<x.\ltlon qUOlil'nt on a I>:.I"C 01 only brge mct ropolitan ;In::IS (I.Q ' l. (·O'T<..'t.'L" for Ihis. 'i1ll1s, agglomL'ralion <X.TU'" " here tilL' ('·oIT<..'(,·tion is mo.<;1 mL'anin~ fll l . Ih:.1[ is, \\ here till' ("h;l n~l' III tht;' 10<'<1lion qllolil'lll from I.Q 10 LQ ' has IX't'1l large, If tlK' ""Iu(' o f I.Q ' IS sm;IIJL'1' 111;111 LQ. I hl' flow o f g(xxh or scn k es is from the larJ.:e ('ily d o wn, \\,~l!'(L jf LQ' h l:trR<..'r 111;111 LQ, lht: 110'" i:- liP" :,rd into IIlL' l<1rg(' cil Y,' Tahle 1 sho"'s IIlL' a\ 'Cfa~L' I.Q for ;11 1 i n d l"~ ­ triL's (,OJl1pu IL'd :lno.~.~ our '11 Ill:t jor I11l'lropolitan area!'> O il <t h a." l' ("o nsisl ing of (11/ {'S. lllL'tropolitan arL':t... and 0 11 a baM: o f o nly m<1 jo r Illl'tmpolilan ;[rL':l .~ cI.Q ' I. T hc Ill'!\Tnt;lgl' (IL'di ll(' o r increasc in IhL' J.Q b "Iso md i{·;lIL'd . S•.X lol':- :11«,.'('l l'(,l hy I hb shirt in till' I.Q has(: . 10 the cx tent of ch:lll~ing thL' " :.,IUL' of LQ by 10 rx:n:ent or Illorl' . arc Ii..tt'd IX' lo\\ Thcy ;m: gn lllpcd hen.: by Ih(' dirc(·tio ll o f Om, . Th i .. lIst shows ho \\' 'hL' typical I:tr~t' 1llt'lropol itall ;IIX.':1 in Ihl' Cnil<,.' d St;lIt'" i .. linh'd to Ihe sll1:1111.'r nWlrOpo !il:lI1 pl.ICcs th:!t lit, Ix:low it in I he urh an hiL'l~m:: h ~. '111b IS Ill(: n o w o f g<xKls ;lnd .'o(,.'IY ia.:!'> het" l'L'n 11K' top lil'r o f Ch:1l1 1 and pl:.lccs hc lo', it in tlK' urhan h iL' r:.lrch~'. l rJ/l'u lY.1 jlo//' 10 lIIaj"" 1Ifc>/m/KHilw/ Cl'I/I(!n~ farm ing; :1J,:fi('· ulturL'. forcsll),. :111£1 fi:-ht"), st'r"in'~ : ('o al mining: oil and gas explor.lIion: railllJ:td tran" I)()!1ati<ln; tfllcking: ('lL'clric, ).,:as, and !'>: lll i t.lr~ ." L'IYicL's: bUikhng lllatC'ri:tis, Econo mic Mcvlt-... - July 1990  Downll'ard flow fmm major metmpoli lOIl 1llL't:11 Illinin~ : w:lIer tr.tnsportation: air II~UlSIXJrt:tt ion : 1r:l nsport;lIion st'f\'1ces: \\'holC'sale tr.lde ; IY.lnking and cfL'<..lil "RcllciC's; :-i(..'(.'uritr ;lnd comn1<xlity d eak"': insur;H1~L' :'Igent... : rt::l1 estalc :<f\'i('t's: holding and in l L'StlllC'nt r Olllpanics: h usiness sen'ices, :.llllllSl'IIK'nls, IlIOlion pi('tllres, Il'gal ~ I'yice.<;, t'dIlGlfinn.. 1.~I'\'ict's. Thl' Ih l of :tll go<xl., :tnd sely ict's is signifi' can tl y lon!!l'r ror the downward Ilow from large 10 smalkr cil ies than \'ire: n~ l's:l. This ind icates the imporl:l ncl ' of agglo llll'J':lIion i n llIa ny urhan ser' \·ic(;'s . Thl' !low of services down Ihl' hi e]~' rc b y cOl1si:-ts la rgely of bu:-i neSs anu fin:mcbl services. jobs and busim.:s.'ot::s cOllllllonl\' asS(X'iaIL"{1 \\'i th 111:l jor urh:tn :lreas ;lild thei r SL'f\'ict' secto r. ExPOrts and imports to other metropolitan areas. T;thk 2 Ibis I\"o-di):tll indusl ril's and shows 11lL' 'I\·eI4lgc: LQ and swndal'd tle" iation o f LQs .. mung I hL' 11 Illclmpolit;tn llfl';ts on a m.a jor llll'lnllxJlil:m baSC'o111t.'Y are 1<1I1k<.'d by ,t:mdaru Cf!Il/C'fY.  The data are '10m I/!$ Butcao 01 Ec(Wl()ffifC AnalysIS. fie-  grotIItI fconomoc /nfomlal/Ofl Syslem o.IIef8l>cC$ I'l f/lc COSI 0I1Mog III7IQtIg Cdlt'S Call bias /he IoeB/1Ot1 quOloenlS bul a eme/III 100II BI /hts ISsue I/ldocated  IMIII IS 8 StTI8I/ J}fOl)IeIn Nolo mal ovemN hvrng cosls I'll" csncel 001 ai/he numefs/Of 0I1he LO fOf S ~I/ic place The base W,1S ffJ(;ompuled uSIIlO AmerICan Chambo< 01 CommefCO ReSfMfChOf', A SSOCIBIIOt1 (AC CRA ) daIS on 11~"l(J COSIS 111 the U  5 melfopoJllanafeas Af/ftf adflJSlmCfl1  only IhfH IOCIII/Ofl quofl/1fliS /0( S 1)'PICa/ CIIy WCfe alfected  somuch Ih8l/11e)'movedby IOperCanJOfmorC BUjlBfage WoJS aflee/cd by 1M f'uQIl concefl/fSlOOfI 01 bt(Jl<ws If! NeY.' YOfIl and by the I1If}II cost oIlM11(J In lhal CI/y The otIlC<  e,l)'  /WOS6(;/Ofs. fl'Il.IS6ta7ISand popellnes. 8£a ~ small S8C/OfS wr/h dlSClOSlJre prob/8ms It! 5()'1')8 CI/IOS and the change 1(1 LO IS hBldef 10 8//nbuto 10 IIvInQ COSI$ a/Ofle None 01 lhe dolloJ shown 1(1 IhI, 8f/1clearcllldfllS1ed foI tw>g costs GIY6fl  .m.n  !he smaI chafIOeS ad/l'S/moMlS we made. and pcsSlbIe fl!SBtV8llOflS 8boullhifl COOJerBge 8ftd ~acy 01 AC, CRA dB/a /he odp5lff1tY1t  was dropped  The ~«IefISNe LO II$6d hefe uses tolal me!'oportan ciJ/7lIOIlQs WI /he base /lie feS/ftC/cd base I$lafge nelfopolo, Isn eartwlgS rhe IIllefpt8iS/OOfI 01 me movement 01 goodS and seMCes snJPI'iI becO'nes 8 /Io<v OOO:'eCtl large and  smBII me/topoillan plltCBS. naf oel_lafge IJlC/lopoIr/8f1 ItnO aN pIM;es SOe GItnetf. K8If. and Mack (1989b) /Of an ex0'8nallOfl  ,  Table 1  Changes In the Value of Location Quotients: Shift from Total Metropolitan Base to a 44.t:ity Base Locatio<> Quotients Seelor  44·Cily Base  Farm Agricultural services, forestry, flsherlas Mining  1.31 1.04 1.1 5 1.45 1.79 2.19 Ul' 1.04 1.00 .97 1.02 1.01 1.13 1.09 1.22 .94 1.00 1.07 .84 .9' 1.05 .97 1.03 1.08 1.02 1.06 1.09 .95 1.02 1.03 1.00 .86 .88 .57 .97 .97 .87 .88 .90 .92 1.04 1.03 .96 .88  Coa' Oil and gas Melal Nonmetal mineral ConltrucUon Manufacturing  Nondurables Durables  T_  fran. pertllUon and public utJlIUea  Railroad  Waler PassengertratlSll Air transport Pipefines Transportallon services Communications Electric, gas, sanitary services Wholesale tr&de Retail trade  Building malerials General merchandise stores Food stores Auto dealers Apparel Furniture Eating and drinking Miscellaneous retail Finance, Insurance, real estate Banking end credit agencies Securily and commodily brokers Insurance carriers Insurance agents Reat estale Combined feal estate and insurance Holding and investment companies Services Hotels Personal services Private households Business services  Percent Change In La  All City Base  .58 .89 1.06 .88 1.60 3.13 .83 1.03 .96 .• 9 1.00 1.07 .96 .96 1.93 .96 1.25 1.1-4 1.06 1.02 .93 1.10 1.02 .97 1.04 1.04 1.01 1.04 1.04 1.02 1.02 1.06 1.21 .• 6 1.05 1.07 1.20 .95 1.07 1.01 .93 1.01 .99 1.06  126.2 16.5 '. 1 64.1 12.1 -30.2 25.9 .5 4.9 9.9 2.4  -5.2 17.8 13.9 -37.0  -2.7 -20.5  -5.3 - 20.6 -4.5 12.9 -12.0  ••  11 .3 -1.1  17 6.9 -5.1 - 1.9 .7 -1.2 - 18.4 -26.9 -33.7 -5.0 - 10.0 - 27.4 -5.1 -15.4 -9.0 12.4 2.2 -2.7 -16.8  (continufld)  6  Federal Reserve Bank or Dallas  Table 1-Continued  Changes In the Value of Location Quotients: Shift from Total Metropolitan Base to a 44-City Base Location Quotients  -"  44-City Base  Auto repair Misceltaneous repair Amusements Motion pictures Health services legal services Educational services Social services Museums Membership organizations Miscellaneous services Government Federal civilian Federal military Siale and local  ue\·iation. lk'){innin)( .It Iht: lOp of Ih~ whit: with Ihe..: .~ma ll esl ~t;l nu<lr<.1 ue\·iation..... \Vhen lilt:., sl'lndard dt:\';al;on is .... mall. and 'ht' an·r..lge of th~ LQ is closc to 1, tht: industry is nonh"sil' and i'> pro\'itled loe'llly with It-w expons or import.'<. In thi!-. I;SI, health s~r";ces, 1.; onslrl1(.:t ion. m;my rel:lil serd<:es, sc\'eml persulllll :md rep.. ;r .<;el"\·ice.... , .mtl whoicsak' t"Jde closely fit Ihi.... dt'.<;<.;ription of loe;11 nonhasic sClyiees pro\"itkd hy Ihl' typital large metropolitan ;1I'e;l in Ihe Uniled St.lles. At [hl' honom of the list in Tablt' 2-\\ h<:rt' lar).\l' \';l ri;lJln: plu\·ide.... dear example .... of goods or seIY;c('s th;ll an; exported anti impol1ed hy nwny dlies-;IJ"t.' water [r..tnspoI1:uion. pipcline~. Illilit.:uy H!SCI"\';ltion.... oil :lIld WI:> t:xplor,lIioll, ("oal mining. ;1Ilt! mct;d mining. lktwcl'n the top and hottom of the list in Tahle 2 an.' sCI"\'ice~ that ;Ift'" prohabl}" mixed. lIlorl' or less widdy exportt'd ~I nd imp0l1l'd hy some cilies. Husincss Sl'IYict:s. amllSClllcnts. and banking af(' in the top h;df of this Illid-r.m~t' and are:: mostly 10GII services. with a small standard Economic  R~vk""-July  1990  1.02 1.02 .89 .45  99  ••• .• 4  .• 9 1.09 .95 92 1.03 1.01 1.31 1.0 1  Percent Change inlQ  All City Base 1.02 1.04 .99 .• 9 .99 1.05 .94 .93 1.01 1.00 1.07  .3 -2.4 -10.9  -49.6  •  - 18.6 - \ 1.2 -5.0 7.• -5.0 -14. 1  .97  •.0  U16 1.01  -4.1 29.5  .94  •. 9  dcviation. But thc stand:ml error b large enough to indicarc Ihat S0111('; ,nlde is likt'ly ;HllOng brge eit i t'~ Flilther down thl' Ii.~t art' Sl;'elOrs such as tnlcking. II:Ul.'ipOn;lt;on .... l'l"\·it·~S. and f..'<:lllc:llion:11 st'n ice~. whidl ;H'l' rrohilhly morl' w iddy tr;lded  but w ith  :t  strong locd <:omponent  ;I~  well.  Four Texas cities These trriGd ex;unples or hasie and l1ol\b;JsiC" sl'nic"cs aern"s -14 major LS. ('ilk's lead u . . 10 :lsk how four T(,;xas citie .... comp:lre with onl' :mother.' Table .i . . ho\\'s Ihe loe;ttion quotients on a ·"\-1-eltr ha.'>l' for Hou....[on. Oil))a!>, S:m Antonio, ;Inu Fort W011h. Fo!to\\'ing Kt:il .md Mad; (19H("1).:, list of ~(,; r\'iCl:-S(';l"t or import:> ;md t:xports \\";\ .... tk'n'lopcd. assuming t'xpoI1." occur if tllc lQ is ).!rt'att'r  ( F()t examples of Slfnil/JI sludles soo t(tJI/ and M.:fCk 00 /ndt. anap%s. GUlS/16f1 ()tl CIeWlland. Md GIknet. Xed and Mack (1987) ()tl e'/lfJs 111 the Tennessee Valley  7  Table 2  Mean and Standard Deviation: Location Quotients on a 44-Clty Base Sector  Average Location Cootient  Standard Deviation Location Quotient  .99  .15 .15 .16  Health services Miscellaneous retait Construction Personal services Eating and oonking Furniture retail Auto dealers Food stores retail Miscellaeous repair Building materials retail Auto repair General merchandise stores Insurance agents State and local government Apparel retait Miscetlaeous services Wholesale trade Business services Amusements Banking and credit agencies Social services Holding and investment companies Membership organizations Legal services Communications Insurance carriers Electric, gas, sanitary services Trucking Private households Transprortation services Nonmetal mineral mining Nondurable manulacturing Hotels Agricultural services, torestry,lisheries Educational services Passenger transit Combined reat estate and insurance Reat estate Durable manulacturing Security and commodity blokers Railroad transportation Federal civilian government Museums Air transport Motion pictures Farm Water transport  1.00 1.04 1.03 1,03  1.02 1.09 1.06 1.02  1.06 1.02 1.02  .97 1.01  .95  .92 .97 .88 .89  .88 .89 .90  .17 .17 .17  .19 .21  .22 .24 .24 .25 .27 .27 .27 .28  30 30 .32  .97  .33 .34 .38 .38 .39 .39 .42  1.05 1.09  .43 .44  .96 .84 1.04  .45  .97  .51 .52 .55  .95 .86 .98  .48 .50  1.04 104 .84 .94 .88 .87 1.02 .57  .59 .5' .62 .80  1. 13 1.01 1.09  .97 .98  1.00  1.10  .56 .56  .86  .45  1.06  1.31 1.22  1.25  1.60  (continued)  •  Federal Reserve Bank of I>alla$  Table 2- Continued  Mean and Standard Deviation: Location Quotients on a 44-City Base Standard Deviation location QuolienJ  Average location Quolienl Pipelines Federal mllilary government Oil and gas mining Coal mining Melal mining  than I.::! ,md import... O(TUr if the LQ i.~ k·ss Ihan Tahle "I .... how... thb list for ~:;Kh city." A.~ diS(.·usS(.·d (;':Irlil·r. in .'>()Illl· GI."<.:S IlIMl·. tedmolo~r. or institlnion;tl dilTerem:l's e,m force "ariation~ in IoGltion quotienf ... Ihm arc nOI indicaI1\T of an t,:XPOlh or :1 II':.Kling p:llfl'rn. On Ih<.:se ~rol1nds. I drorp~.:d from Ihe l ist o f pOfentially tmdt:(l ... t',,·in: ... tht· follo\\·inW r:Is....engl'f lI':.msil: motion pi(·fUrl· ... ; prin l<.: hOll.'>l'hold employment: combilll'{1 rt'al estate. kga!. :lntl insur.mct' onin:s: and social .... <.·Iyio:s. In general. r did not S<'T t h<.·sc d .. ssilkali()n .~ playjn~ :1 I1wjor rolt' in inlL'rdlr 11~lde. :md I fell tli:lt their \·ari:mcl' .~Iemm<,'d from other ...ource.... For example. the Soufh"'C'sft'rn p rdt'fl'IKe for tht· pr;\'atl' ;mlo explain... the 10\\ l<.lcui(l1l qWllil't11s Il)!" pass<,·n).!<.T tran .... polt:llill1l in ;.11 of Ihl's<.· dlil'~ In "'(1111<.' cases. il1lmediatl"ly diS(.·t·rnin!l pat· tt'rn~ of t'xporh or imrOlb in M;'(·IOI~ is c1i1TInlit. T\\'I, .~lIdl l'X;II11pk::. ;Ift· f(x){l ... tort's lInd uK'lllber· sh ip organizations. ll{)\\·e'·l'r. in retail .'ol'dors. lllajor warehousing EK"i l ilit·.... or he:ldquarter... or chain sIOl't·... can gi"e ri~ 10 reJ.:ional t"q){)lt:.. Hotel:. and re ...laur;.mL'" export' ;:1 touriSlll. conn'n(ion.... and hu~ines ... 1r:1\"(:,I; nalion:.1 or rt..'gion,,1 hl':ldqllartt·r... of Illedkal. t'l1!linl·l..'ring. or publil: Sl'lyi('·e organi/.;ltiolls bel'Ollll' l'xports of lll<.:mbt'rship or~ani7.atiol1s Intt'rprclinn (If tlK'se n>t'1TIdenL~ rL>(luire ... great GU't'. esp(."Cially \\ hen "pplicd 10 case ... ,1:spt'dfi(' ;l.~ Ihe L' i lil'~ I haH' L'ho:'f:I1. Still. rill' bmau pattern oj' sf:l'\il'v·.~t'(,"lor sll'l'ngths des('"l'ibt,tI hl'rf: l.'Onrorrll.~ 10 image.... held hy knowll'dgeabll' oil·  ox  F.cono m ic Review-July 1990  1.07 1.3 1 1.79 1.45 2.19  2.70 2.72 3.44 4.74 5.45  :,c1'\'crs o f tilt: citit:s. These resu lts possibly ('Olll:lin .... 01l1e new inSight as \\'t:II. Housto n . iiouston b an oil :md natural ga~ center for Ihe world. :mtl Ih,' 1(x.~lIi()n quulll'nts for oil and ga ... mining (16.'75) ;lIld pipl'lin(.' 11-.IIlS, P0rl;llio n (9 I I) "i\'idly illlL~t l~lIl' Ihis re)in!. Of Iht' l Op filil'en pllhlidy hdd fi nns llt.:adqu;u1ert·d i n lIouston (rankt-o by anOlt;ll n..·'·(.·nuf:.~ III IQHH). 11I0l' art' relaK"d 10 oil :Inel gas: Tl·nnel'o. Tht' Coast:.1 Corpomtinn. Enron. J.rondeJl Pl'tl1X'hemiGIL l':tn ll:tndk Eastern. TI~ln.~l.'o . Bakl' r- H lI~ h t.: .... Pt'n nian P: 1I1I1CI~. alld I't'nnzoil (Schadt"w;l1d 19H9). Mud, of Houston 's appal\.·llt strt.'n~th in ....<.:"·il'l"... stl'Ill.~ from oil :tnd g .. ~ ;IS wl'll Tnl(·kill).!. Ir~m,~p()rl:tlion selylt'(.·.... ilnd ":Iter lJ~IIlSpOfl;lfion :11\.' alillosl l.'t"rI:tinly linked 10 Ihl' Port of Houston. B;ISl.'d on tonnage. it i .. the largc~1 (l.S pon. and pCtrodll·lllit~lr.... ;lliel rcfin"d enl'rg~· prcxlun ... (;llonJ.: \yith awkuitur:11 products) ar<.· tht· top t.:XPOll ... from tIlt: pCl11. T he sll'l'ngth o f Ihe miscdlanl'()ll .~ :'l'l'\'in:s C:llq::ory n:sult:. al k. ,,~t in P;1I1 from lil(' 1;11').!t! (.'on"'Irunion cnJ.:int'cring firm ... in r IOllston sll(h ;IS Fluor D;mici. Bnmn :lnd Rool. ,mel M.W Kellogg. HOWe.lon is :I national :tnel !lloh:d n::nICI' lor t iK' design ;lntl t'tlll.~trllLli(1Il .)( hl !'gt' p<.:tfOchcIllil'a l and rdining pbnts. JI(llistoJ\·.~ Jarw~' numher of  Tho 0 8 a1l(/ I 2 are IIr/JJtra!),clJtoll pOUlts, selected on /he  baSIS of oxpariencc  In  applying tile mellJOdoloOy Thus, 8  "SIllS"' siane/dfd deVl/lrlO(llsless than 0 2r1l8bsoiule va/uc  •  Table 3  location Quotients for Malor Metropolitan Areas In Texas on a 44-Clty Base Secto,  Houston  Farm Agrlculturalaervlce., forutry, fisheries Mining Coal Oil and gas Metal Nonmetal milleral Construction Manufacturing Nondurables Ourables Transportation and public utilities Railroad Trucking Water Passenget transit Air Iransport Pipelines Transportation services  .76 .87 12.64 .67 16.75 n.a. 1.11 1.33 .74 1.22  .4,  1.29 1.16 1.23 1.86 .42 1.06 9.11 1.85 .83 2.29 1.23 1.03  Communications Electric, gas, sanitary S8l'Vices Whole.... trilde Retail tl"llde Building materials General merchandise stores Food slOles Aulo dealers Apparel Furniture Eating and drinking Miscellaneous retai' Finance, Insurance, real estale Banking and credit agencies Security and commodity brokers Insurance carriers Insurance agents Real estate Combined real estate and insurance Holding and Investment companies Services Hotels Personal services Private households Business services Auto repair Miscellaneous repair Amusements Motion p!CIures Health services  .91 .97 1.12 1.16 1.23  .86 1.05 .86 .• 3 .• 3 .51 .73 1.15 1.65 .58 1.52 1.06 .80 1.1 7 1.56 1.05 .98 1.11 .82 .10 1.03  Dallas  .61 .88 4.92  .5' 6.1 4 n.a. n.a. 1.27 t .13 .96 1.21 1.39  .4.  n.a. n.a. .72 2.60 n.a. 1.07 1.29 1.07 1.55 1.26 1.04 1.37 1.32 1.24 1.19 1.14 1.25 1.31 1.34 1.21 .64 1.56 1.28 2.80 .48 1.55 1.03 1.43 1.32 1.21 1.29 1.18 .99 .76 .30 .85  San Antonio Fort Worth .66 .70 1.4<1 0.0 1.57 n.a. n.a. 1.13 .45 .61 .37 .71 1.14 .8' 0.00 .43 .42  u .33 1.08 .53 .64 1.2() 1.33 1.06 1.53 1.23 1.13 .82 1.44  .77 .83 .80 .19 1.50 .79 1.04 .22 .74 .80 1.21 1.26 .94  .68 1.14 .81 .68 n.a. 1.00  .6' .76 1.73 n.a.' 2.24 n.8. M  1.04 1.22 .7' 1.44 .'2 2.83 .85 n.a. .52 1.98 n.a. .46 .56 .72  .80 1.08 1.04 .96 1.13 1.18 .68 1.48 1.19 .8 7 .52 .60 .15  .4. .86  .6' 1.04 .70 .66 .66  1.32 1.01 .53 .97 .73 .99 .08 .82  (continued)  10  Federal Reserve Bank of Dallas  Table 3-Continued  Location Quotients for Major Metropolitan Areas In Texas on a 44·City Base Se<1O<  Legal services Educalional services Social services Museums MemberShip organizations Miscellaneous services Governm enl Federal civilian Federal military State and local  Houslon  Oallas  1.33  1.11  .87 .60 .81 .83 1.27 .83 .61 .23 .98  .55 .56 .59 .81 1.03 .75 .7• .25 .83  San Anlonlo Fort Worth  .73 .7• .77 n.a. .78 .54 1.94 2.85 5.94 1.14  .45 .44 .51 1.52 .75 .55 .74 .74 .81 .74  "n.a.-I"IOI available  t'nwnt't'ring jobs. and Iht· apparent :-trength o f construction in theS<.· da!:1 O.Q - 1.33). ean be aurihutt....d to tbe petrochclllical con:-lruction hoolll along, tilt' Texa!'> Gulf Coa:-l thai I.x:gan in 1987. E\'t'n IIOliSlon's ilpp:m.'nt CXP<:)11.; o f real est:ut..·. holdinR and inn!:-tment comp:mies. and legal :-t..'I"\·icc:-; arc probably liL'd to the \'olume of oil ,lOci 1(:1:-; proie(1 ~ fimmct..·d through tht..· city.DaJ.las Dalla:-. in t'ontl'astto Iiouston's derx:ndt:nct' on petroleum. i:-. a more widely di\'ersil1t'd cit)" with a mOljor rok 10 nat ional ;I nd regional distrihu tion. A.~ in Il ou,~t{)n, oil and ga:-. i:-. ;111 important busincss in )):111:1:-'. as tll(' LQ of 6.14 for oil :lIld !p!> mining indiclll:S. Unlike.: I iouston, the.: 11l0st :-trikinF a s pt..·ct.~ of the Oa llilS sel"\'jet: sector are il.~ .~trt·ngt h:-. in wholesale trade and in a range of rt.·tail .'t/:::"I"\·iCl:s Ih"l .. rt... Ilonmilly purely IOGII acti\'itit::1):llb:-. b a l't:ntt'r for rcginnal and national hcadquartt:rs. anti Tcx;l~ In.'>truments. llV,.1 C. Pt'nncy, OrC~Sl.'r Indll.~trit..·~. Southb nd Corporalion. Kimberly Ct:.rk . Ekl·tronic Oat3 Systems lEDS). and Frito-L:I)" iHu.~tr.lte the divcrsit}" that e xi sl~ among the t~'JX!s of indu"t ries with head· quarters in the city. Dall:l.~ i:- physically located 10 .~I"\"t..· a:- a dislribution hub for tht'" Scmtllwt:st and rur ~111.~ o f Ihe ~ Iid \\'t..'.-;I a~ well. Tht'" bugt' World Tr.ldt.· Cenlt:r i.~ tlK" mOM conspit"lloliS !>ign of the t'ity":- hroad fOIl..' in national d ist rihllt io n. TIlt' IOl';ltion qllotlt'"nt:- for Dallas show exEcQ nomk Kevlew-July 1990  ports in sales in industries ~"Uch as food stores, eating and drinking estahlishments. dep':lItment store:" automohile dealerships. and misccllant::uus retail stores. D:lllas' strength in eating and drinking places, hotels, and personal sen'ices tics in to its largc numl.x:r of cOIl\'cntion visito~ :mel other nusincss tr.a\"dcrs to tht.." city. The large DallasFOil Worth Airport and Amcri<.<m Airlines headquarters in tbe city .~ I;tnd (lut \'cry strongly ( LQ ., 2.60l. I)all .. s s how.~ a much stronger financial sector than I (Ollston, with (."xports of banking, real est,lIe. and insuranl·t'. San Antonio a n d Fort Worth. These smaller m~troIX) l i t an a re:1s, in conlr..lst to Dall;ls ;Ind )Iollston. have fe"'er selyicc-secto r exports ,md far more imp<1I1)'. In S.. n Antollio, th~ large 10C:llioll quotients for the fedel"lll government  . One mpotlanl nonenergy IJlPfJ'1 J IXpecled 10 see. and whtcIl does fIOI appear .n ",,$ dal. 1$ mecJK:aJ expotl$ The Tex.s MedIC" Cenler .n Houslon ~ 50.()()() OIO'1<ers WIll! tIfIOIfllIr IO.()()() sl\.ldll'llS II $8f\I&$ c/Ienrs Irom movnd !he WOf"I(j In rhts ctaf• • Ihls #lXpotl falls ..elm 10 lItO (JIOb-  /oms (I) 8 reiahWiy.."at< local h8aIlh sec/Of III HovsIon MechclII Cenler. and (2) emp!Oy8eS" al the CetllfIf tePfftSenl #I rntX 01 (XMJII and pubI1c 1fIS/,/uIJO(1S  aosrt !rom lhe  Wlm pubftc emp/OyeIs (such .s18ClJlfy ,'Ihe SlalemediCal. denIal. and nut"$U1Q schools) 8{)De81f1fIg If! file puoIK; pay.  ro#s Aga.n. /his,s one mote ell/!lmple 01 ffI8 care rflQUlled the 1II/6<(XII"JO(I 01 OI.X r8StJ/Is  If"!  11  Table 4  Servlce--Sector Exports and Imports In Major Texas Metropolitan Areas, 1981 Hou.ton  Exports:  Imports :  Trucking; Pipelines Water transportation Transportation services Wholesale trade; Retail apparel Electric, gas, and sanilary services Real estate; legal services Holding and Inveslment companies Miscellaneous services  Security and commodity bfokerage Insurance carriers Federal civilian government Federal military governmenl  DIU••  Air Iransportation; Communicalions General merchandise retail Wholesale Irade: Food slores Eating and drinking establishments Auto dealers; Miscellaneous retail Banking and credit agencies Insurance carriers; Insurance agents Holding and investment companies Real 8slale; Hotels Personal services; Business services  Railroad transportation Security and commodity brokerage Amusements Social services Privale educalioo services Muset.lms Federal civilian government Federal mifilary government  Fort Worth  Railroad transportation Air lIansportalion RelaW furniture  Transportation services Communlcalions Electric, gas, and sanitary services Security and commodity brokers Banking and credit agencies Insurance carriers; Real estate Retail apparel Miscellaneous repair Holding and investmenl companies Holels: Business services Private educational services Membership organizations Miscellaneous services Federal civilian governmenl Slale and Iocaf governmenl  Pe~nalsennces  Muset.lrns  San Antonio  12  Building malerials Eating and drinking eslablishments Food stores Auto dealers Personal services Insurance carriers Hotels Federal military government Federal av~ian governrnent  Air Iransportation; Transportalion services Electric, gas, and sanitary services Miscellaneous relaillfade Wholesale trade Insurance agenls Security and commodity brokerage Holding and investment companies Business services; l89al services Amusements Private educational services MemberShip organizations Miscellaneous services  I'c:deral  Re5erv~  Bank or Dallas  reOecl the ciIY',s dl'l'>Cndcncc on the porrolls of FI. $;un I Tou,slon and four major All' Force fadlitie..... F.xrX:n1S of personal scl"\"ice:-., hotd..... <lod e;lting and drinking places ;Ift: related to tourism . St'r\'iCC-SL-ctOr exports from Fon \\"iorth <Ill.: ('onnned to retail furniture. personal :>.Cl"\·i<.:e.'>. J)1USClJln..... and a strong tr<lnsport;lIion Sl."'<.:tor (mil roads ami air trdnsportationl. Uoth <.:itic.' impOit a widc "Lnge of sen ices. especially financial and husim:.ss .ser\"lct:s. Wilbur 'l1\ompson (1%<;) propo~d four staJ(t' . . of urb'LI\ growth: ( I) thc expansion of an expo]"I-lxlsed cl'onolllY k:d by one expoll : (2) thl' rise of on expol1 complex. as olh(.·r compl..'ting exp0I1:- den:lop; (3) Illl' achic\'ement of economic maturity :IS n:gional manuf'l(.·turin).\ displ:lCes imports: and (4) thl..' dc\"(.' lopmcnt of a rt'gional J1ll..'tropolis with SCITil..·c-:.<:l10r cxfX)O.~. Any st:lp;e theol), of this kinu b an o\'cr:-;implifk:ltion, hut thi .... dc.~ription .... txm .... u,cful in li).:lu of find11l~!'> for these rour dtie!'>. Dallas :lIld l lou:.. ton han: achie\~1 the final ....1<L).:l..'-.st;ItUS a.' it rc).:ional melropoli!>. pro\'iding serviLe:. to other CltIC.... throu).:hout the Southwcst and Cnited St;lIe..... San Antoni o and Fon \X"oI1h. on Ihl..' other hand. n:maln .11 an earlier :.I,,).:t' of dt'\·clopmelll. '111e~ ;Ire proh"hlr malure in tht' st:n-"t' of :o;la).:1..' 3. hut ;.IS }'l't thl..'}· show only limitt'd st'J"\'icc-s\"'ctor e:<pOlls. Conclusions  ThL, ;1l"ti(.;[(.' dCJ\lo n .~tr:Jt(.::-. :I new war 10 <l i\'ide the tlll..,t1"Opolilan sl;'l"\"i<.:(.: s<.:ctor inlo h;l:-.ic and nonb;Jsi(" indll ....l ries. By :.liRhtly altl·rin).: ;\ !'>tandard meaSLlfl' of <.:xp0rt potential- the location quotient- I \\as ahk' to I.~()l:itt' exporlS thaI gl'n(;'l":.ltc growth in lll:I]Or lHl'tropo[itan ;11\.':1:-' . Fll11her. possiblc ("(lOfusion is diminat(;:d hct\\ l.."'t:n the ;Ibility of 11)0S[ urh;tI\ :\1"(; :1:. to .'>C1"\·C the towns <lnd small cilies 111 t[wir hinterland ;lntl Iheir ahilitr to export to other lar).:l..' llK'll"Opolltan afl..':ls thmughout the Umted Sl;Itl.." . Somt' l... u"t'flll 111tl'rprl'lation of the dat" I.'" required . '" with ;IIlY l'l"(momi( methodology. hut ;I hroad picllIrc of se1"\ ICl..·--.eClor :.tn:nglh ... ;tnt! \\·eaknesM:':. t'lller).:es deilrly from 0(.11' case ... tuur of four Texa .... cilies The (·omlxu"i....on of long-time rh 'ab Houston and Dalla... b parti<:tdarlr 1I1tercsling. Houston profe.s.,es to lx' the oil l":.lpil:11 of the nation, and it:o; Eco no mic Review-J uly 1990  gloh;!1 opcr.llions in oil :lnd p<.:tfOc.:hemic.tls ~in' it . . trong claim to Ihat tille. A look at otlK'r m;ljor J)1<.'lropoliwn arells Ilwt might make simil"r daims -Dallas. Fort Worth. Ik:n\"o,~f. Of New Orlean:-.-sl1<)\\:' no re:11 cllllllc nger. !\:e\\' Orleans has the sl"'('ond highes( Ioc;nion quotient il1 oil anu p;a:-. Illinlllg (9.h~J {'oll1parcd 10 lIow:(ol1 (16.-"). Howe\·er. !\:l..'\\" Orl~ln.. · l·aming::. f!Dm oil and ga.~ mining are on ly 5631 million. or 20 fX'rcent of H(luston·:. S.:\ '] hill ion. In rar!. Dalla~ comes do.~­ esl JI1 thi.: unllar \';due of e:1I"J\ings from oil and gas '''itll S 1 hil lion. Add Iioliston \ t.'xtensi,·e refining and petmdll'mic;11 ope ratiol1~ and thl..· f:let th,1I no major mellupolit;m area in Ihe nation rh'a b it, pipeline oper;Itions. anu Jlouston takes thr.: title of national t'ner).:}" l..·apit:l l \'eIY e"sil~. At the sallie time. I).l llas CI\1(.'r).\e:. as Ihe di.... tribution and l'inandal capital of thl' Southwest. exporting tlle~ S\..·lyice.' widclr to other major cities. Dallas' only lIl:ljor mctropolitan rh'als (hal lie Ilt:t'" cen thl' t\li~si.ssippi Ri\·<.'r and Calirornia and th.1t h;l\e a s(mll).: Southwcstern oriental ion would ht' Houslon. IX-m·ef. l'hOl.'nix. New Orlc,m..... and Salt l:lkt' City. All of thl..'sc..' citil'S dCmOJ1SIr.:lIe some t'xfXm stn:n).:th in ret:liling Of l'in:.mdal .'>Cr\ · ict~..... hul nonl..· .,how tbe wi(k.... pre:ld :.Ircngth :lIld diH'r., itY in ~r...icc:. of the D:lI1" . . economy. Oalla . . <.::m 'K<"llr.llcly Ix' l..·al1ed Ihe fl'gionall..'{-onolllic l '''piwl of the Soutl1\>.cst. The rh·.dry bet\\"t'l'n Ilotlsl\ln and ]):lll:Is is I OI1~ ;lnd l..'nuurinR. and I wi ll not try to de('ide here w lwthcr it b heUl..·r 10 he l he nation's encrgy c;Jpit;d or the r{'gion;J1 ('<..'onomic capil:ll. But this pcrspt'l.."til·c on the L\\,O dt ie:-. pro\·ide.... insight into their rok in Ihl..' SOlllhwe.'ll·rn e(.·oI10my. Both l'ilie!'> h;I\'l' ;111 extremely pmYt'rfll l l..·collomic hasl..· th"t. despitl..' tlK' ri",dry, I.... Ihrc:Henl'(l only langenli:llly hy thl.' other city. There is o\"l.:rlap in economic funcl/on.' . of COU]":;l' . I mentioned Dallas' role in oil and ).\<1 ..... and throll1-thoUI the 198<h Ilou.... ton h.ls ",oLl).:ht di\ l· .....iI'il. ltion of it.' export ha.... e ;1\\ a~ fJl)m oil :lIld ).:;1.' . HO\\'C\'cr, II J~ unlikely that t\\'o m;Jjol" nK'lropolitan arc,,!'> the size of D:llla.' anti Houston t:ould cxi.'>t for m::lrlr l~ ye:u":-. :lI1d Ie....' th;m 2')0 milt':. .. JXIn if lhey f(..pfl.... !'.'Cntl"'"<.l a hmd:ullentill thrt';!t to l'ach othe(!'> hasic export!'>. A s(:conti int~rC~ling ('ol11parison i., Ihat of ])all:I'> .11,<1 llouston 10 the smaller S;.1Il Antonio and FOl1 WOIlh . Unlll of tl\(.'Sl' sl11:1llcr I11cl ropoli.  "  tom areas are home to 1.3 m illion people, yel I found a limiled abililY 10 achie\'e service-seclor exports. T~hniC".. 1 and agglomemti\·e econom ies in Ihe scn:ice e<.-onom y are depleted quick ly 'IS  we mo\'e down the urban 1:ldder size.  b}r  population  Appendix Methodology for Identifying Service-Sector Exports  This appendix provides further detail on methodology described in the body of the article. The key features of the methodology discussed here are : (1) that the nonbasic location quotient is 1; (2) that agglomerative economies can prevent the nonbasic La from assuming a value of 1; (3) that systematic shifts in the base of the La can correct for urban agglomeration ; and (4) that shifts in the base of the La also contain information about the industries where agglomeration is important.  y 00 = regionwide earnings, basic and non· basic  Some algebraic manipulation quickly shows that:  (1) The nonbasic location quotient is 1as long as we regard al/ places as independent of each other and ignore the urban hierarchy. As in the text, assume that basic earnings (b) vary widely from place to place, and that nonbasic earnings (n) in industry i are proportional to that of the economic base.  Then the location quotient is always 1 for all nonbasic industries and in all places. This can be quickly seen by noting that: Yo; - total earnings , basic and nonbasic, in place i  :b/(l+ i,.,  a/ )  (2) Once we recognize the urban hierarchy. nonbasic focation quotients for large metro· politan areas may be farger than 1. Because of economies of agglomera· tion.large cities al the top of the urban hierar· chy will typically provide services to their hinterland, that is. to the smaller places that are close by and lie below them in the hierarchy. Gilmer. Keil. and Mack (1989a) show that a central city with earnings 1. . and a hinterland that consumes product i." partly produced locally with employment YiH and partly imported from the central city. will have a nonbasic LQ that is greater than 1.  no ,. total nonbasic earnings, regionwide in industry i  Federal Rcscrve Bank of Dallas  (3) Tocorr9Ct foragglomeration, a simp/ecorrection is available in the form of a new location quotient computed on a different base. Economies of agglomeration in the urban hierarchy lead to nonbasic location quotients greater than 1. Keil and Mack (1986) argue that the way out of this dilemma is to carefully distinguish exports from a city to its own hinterland (in other words, those exports driven by agglomeration) from its exports to other major metropolitan areas or to their hinterlands. To remove the impact of agglomerative exports from the La , they shift the base of the LQ to a metropolitan or urban base. The adjusted La is designated La' in the text, and the formula is shown in the body of the article. Keil and Mack claim the shift in base acts as a filter, removing from the La the influence of lower-order places on those cities that stand at the top of the central place hierarchy. Their argument hinges on how the metropolitan LQ changes its value when the base is shifted. They show: ALO  II:  LO.. - LO '..  . k{LO. (LO, -l)} , where L0lu and L0ir are the location quotients for industry i in metropolitan and rural areas, respectively. Given k > 0 (and they show this is almost always the case), the sign of ALa depends  Economic Review -July 1990  strictly on whether La. <, >, or ,., 1. The case of urban agglomeration implies La.. <1 . Thus, the typical urban La subject to agglomeration will be larger than 1, and La' will be smaller and fall back toward a value of 1. Indeed, Gilmer, Keil, and Mack(1989a) show that in a Christaller-losch central place system , this new La shifts back to a value of exactly 1. The inflation of nonbasic LOs by shipments to smaller places induced by agglomeration is eliminated perfectly by the shift in base.  (4) The direction of the flow of goods and services through the urban hierarchy is indicated by the shift in base of the location quotient to L0 ', As explained above, if La. is less than 1 , agglomeration in urban areas is indicated and a downward flow prevails from large to small c ities. If La~ is greater than 1, the dominant flow is upward from small to large places, and the La ' for the urban area will increase when measured on the large metropolitan base. This upward flow is not related to agglomeration but to comparative advantage of rural areas in industries such as agriculture and mining . The upward flow should be regarded as an export from small areas to large ones in the same sense as we otherwise identified exports from one large metropolitan area to another (see Gilmer, Keil, and Mack 1989b).  "  Refrren«s AIl1t'rit-:lll ChamlJt;'r of Comlllt'l"Ce l{ese:l fdl(..'rs As."oci;l t ion (19H~). The ACCHA CuSf qf '-it'iI/X Index (I.o11is,·ille. Kr.: ACCHAI. Andrt'Ws. Ridwrd 13_ (1953-55 I. -hledl;U,ics of lhe Url>an E(:onomic B<I.'>t:.- P'lrt~ I-II. I.L/Ild Eco1I0l1lics 19:161-6- . 16~. 3-1}-50: 30:52-60.  164- 72.260-69.309-19: 31 :4''-53. I·H -55. 2-15- ,>6..~lJ-"'1.  H~s(.' lye Bank of ECOllomic Rl'I'/('I/', QU:II1er 3.2-15.  lhe FOlHlh District." Ft:li(: r.d C I ~n~l:Ind  I\l'il. Stank'}' R, and Richard S. Mack (1986" -Identifying Export Ptxl'nli;11 in the SClyice Set:lOr.- Groll'lb (lild Cbal/ge 17(April>; l-IO. l\inM. It!:-1ie J (19f1'11. Cel/fml Pla(;e 1bl'Oly<Be\'-  t'rly Hills. Calif.: S,lgc Puhlicalion:-, Inc),  Bep:fS. \\:'i1Ii:lIll B...md !'.lichael J Ah·jm.' (19851. "Export SClyices i n Postindustrial ~O(.:ie[r.­ Rc~ional !X'ienl'''' A...."(Kia[icm 1'00JX'f no. 5-' :B - l"i.  Leislr;lz. Lany F.. ;md Sle\'en H. Murdock (1981), 1lJ(' S(x:il)('(:mwmic ImpaCI of Resource Velle/opn/{'I/I: .lIe/hodsftJr ASSl!SSfI1elfl (Uoulder. Colo.: \X'C.<;I\·!t'\\· Press, 1m:.).  Gilllll;:'r, Roht"ll \\'., Stanlcy R Kei l. I{idlard S. ~lal'k t 19W'I. "E.xport I'oll;:'n[ial of St!",'ices in Ihe Tcnnessl"t' \'aller:' Rexiol/a/ Science Per-  fl. leyer. J),I\'id It (1977), -Agglomer.llion EconoI11les <lnd L·rha n-Indu."trhtl Growlh: A Clarification :md Ik"iew of COIKCrt.... ~ Regiollal Sd('lice Pe~/)(~til,(':'i 7( I ):80-92.  s/X,.,:fll'tiS. P( I ): I R-.H. _ _ _ 0  _ _ _ •  and _ __ (1 989;11, -TIll' LOl'ol-  lion Quolienl and Ct:nll,11 Plan: Thl,(JIY." Drati !\lanll~'ripl . . :lIld  ( 19H9hl. -Thl: X-fYit·c  St..,,(·tor in .. 11ie .....Irchr of Hur.11 1>1:10.::-.: POIl'l1[ial for F..xpnrt ACIi\ i1r: hm(/ li:(Jl/O/JIit.;,\~ 65IALI~ustl:  I.isIS, ~  f/ollst0I1IJIISil/ess'/ollrlw/19 (Decem-  ht'r 251. SI;mbal·k. ,1r., Thoma.... " I.. and Thierry J !'\oyclle ( 19H11. Cilie.~ ill 1'nJllsiliOll (Totow:l. I\'.J.: AI I:mhcld . Osmun and Co.'.  21 "7_T.  Groshell. Erica L (IWP). "Can Sen·in;... Be a SOUIH' of Export-l.ed Gro\\'lh~ E\ itlL'ncl' From  .6  Sdl:ldt'\\'akl. Bill. ed. (1 9K9). "The 1990 Book of  ·1110mp,"on. Wilhltr (1965), A PIT!jace 10 l'rixm /:"('UII(llIIics (Ii;lltimo re: The .Johns Hopkins Pres... "  Federal Reserve Bank of Dallas  Keith R, Phillips Economlsl Federal Reserve Bank 01 Dallas  The Texas Index of Leading Economic Indicators: ARevision and Further Evaluation n the ,1 uly 19HH i!'Wllomi<' Nedell'. I presen ted a compositc i n<i('x of IC:lding L'conomit' indicator... for Texa .. ('n .\). Tht, index pl'O\'ed to IX' ;t gexxi prediL·tor of tllrning 1'X)lIlt:-. in Iht' stalt' (,-'t;ollomy, fkcellt strllt:lllr;11 changes in the Tex;I:. t'('OllOlll)' :md tilt;' ;L\':lii:lhility of nt:w d;lIa han: made il IX)ssihk to ('onsl rllCI :L new. impmn,-"tl index. Tht' new leading index inc()qX)I~I\t:S IWO ch:lnges first, th ...• weight gin'n Iht' two eneq.:y ,·;triahles in the index h:I.' 1)1..'(.' 11 n..'dunxl hy Imlf This <:hang\! rencl.:ls l.....(.·cnt l'L"SCilrch, such as fomby ;Lnd IJirs<.:h l ~rg t 19~9). showing that th<:.· T cxas economy ha.' hccol1l1..· ks .. dt:"pcndl'nt on tht;' IC'1lt:"rgr indusu')'. Tht: sL'cond <:h:lng\.' In Ihl' 1l'ildinR imit:x ~',:lS 10 im:orpor.lll· tilt:" 111:\\'ly a\'ailabll: <lma on Tt!X:lS ~xports hy t'oun tl'}' of dt!~ti n a t ion. Wilh [he~e data. I (·o n.~tnlt'led a new Texa~ Ir.ldl!-\\'dghlcd nLiuc o r lile doll.Lf ;l.~ a sllh~tilli t e fo r ;1 kss din:<.:tlr illeasll red Texas \"allil' of the dollar. The nl'\\' T~xas leading inde:< (~TLI) !l1m·t:s \'crr '''i1l1ihLrly 10 tht: o rigin:11 indt'x In fact, the Ilirning points in Iht:" t\\'o indexl's match almosl ex:u':l ly. Ne'·enhdeS!>. till" KTLI ha~ done a lX'tlt:T job of pn:dit·ting mOVCIll\.'nls in the Tex;ls eC()Ilomr. Till' index ;llso lIPPC;Lrs to forehado\\' !>iW nil'k:lnt ch;lIlg~s in thl' r.lIl' of ct:onomic growth. not Illerdr contr,Ll·tions Of upturns E\·\.'I)· month, .1:' difT\.'r\.'m .. go.::ncks Tc\'iSt: thl! Lompom:nt' of the \TI.I. Iht: "'1'1.1 is [\.'\·ist:d for tltl' prl'(.'l'lling SI,.' \·en momhs, ] :m:liyzl'(! tht: "'ffU 10 Sl:l' how monthly rt.·\·i:.ions would hiL\'l' afft.'{·lI.!d It, pl:rfofm:tIK\.' In thl' ...hon s.:unpli: period, monthly rc, 'isions in tilt:" "'TIJ wcre small. :md 1111.: prdi11lin:lly t:stimales were unbbsed :1Ilt!  I  t-~o nomlc  KCYlc w -  July 1990  efficient predi(·tOl"s of the fi n:1 1 index \·aim;s. One conduslon drawn from thcse resul t" is I hal thl' user of tht., :--rru C:IO he (,xmfidenl thai Iht' signal.~ gin:n by the early eslim;lIcs of till: ~T I.I \\ ill nOI ch:IO~t' si).tnifkantly as Iht:" index is re\·iSL'd.  Developing a n ew Texas lead ing index Ahhough tht' 'I'll W:I:-. sensiti\'t;' to (.·hangl's in the Tex:!:. ct:onolll}'. it is appropriatl' 10 lInpron' the index \\'hL!n neW indie.lloJ'1. tx.'come a\':tibble and wht:n thl' t:conom }, changl-'S. St:n:-ml changl·., wert! considcr...'{1 and il1lrl~men t cd. Table ] :.ho",.' tht:" l'omp<>nents and \\'t:ight, uSl:d in both the o ld and thL' nl'\\' Tex;I .~ leading indexes. On\.' :ldju.,llllelll ,,'as to t'onstl1lct ,I mon.' Girl'clly 1ll(!:lSllr\.'d ']'exas lrad~ - \\'eigh l ed intcrn;Ltion:LI ext'llangc rat('. Thl' Tt!xas doll:ll' index originally used in thl' T U ~\·;lS compu tl'd by fir~1 Cilklllllling induslly-specific mClIsures of Ihe dol , lar. based on national tr.lde by inuu."uy and ('OU Il tlY. ;lnd tht:n weigh ting these measure.:s h y Ih<: lIuportanc\.' of the \';uious industries 10 thl' T<:x:l." <:conomy, One problem wilh this type of dollar mdt:"x IS it., di.~rt!g;Lrd for geogl"phic :lnd cultural  I would lIIIe /0 IfI8Ill< JohIl K HIlI TIlomas 8 ':omby Wollam C Gruben ¥Id SlepIWr1 P A 8rQlm lot hcIpIuI comments 1/Wll8lsogrlll4!fuI/OJohIlJ Soorl1r'108IldDAnnM Ozmenl  lot Ih8IIlescsrctr /lSsrsllJflCe  The nmy TelC/IS 1e<'Jdlng /flOe" IS IIvar/abie tTIOrl/tlly. wrthotJl Ch/llge. by wrrllllg K(IIlh R PhfIIIps Research Oepatlment Fecleraf Reserve BanI< 01 DallaS. &alPOfl K Dallas. Texas  15222  "  Table 1  Variables Used In the Texas Indexes of Leading Economic Indicators  Variable  Weight Original New index index  Texas Average weekly hours 01 production workers in manufacturing Help wanted index Real Texas?? stock market index New unemployment compensation claims (inverted) Real retail sales (three-month moving average) Number of well permits issued Real price of crude oil  1.03 1.05 1.02 1.03 .97 1.00 1.00  1.03 1.05 1.02 1.03 .97 .50 .50  .98  .98  .92  .92  NatIonal BEA index of leading economic indicators International Texas trade -weighted real value 01 the dollar (inverted)'  ' Measured diflerenlly in the new index. See Ihe text for further explanation. NOTE : In computing Ihe original index. the weights are divided by 9: in computing the new index. Ihe weighls are divided by 8.  ties to ~kxico: Tcxa.s mar tnlde Ill uch more \\"ith tilat cUlInln.· than the statt'S industry struct ure sllggesL..;. Tex<Js exports by cOll ntry of destination. released by Ihe Foreign T"lde Diyision of the UlIl·ea u of the Census. l·.S. De p.H"tlllent of Comllle rce. ~dll)\\·ed me 10 caku late a mort;' direct measure of the Texas t;'xeh:mgc rate.' Tht;' index  The exp<xt dala came trom a secondary source Texas Department 01 Commerce (19$9) Allhougll tile data are rep<xled byoriginot movemenl andnotorrgmol production. tile slalls/real analysrs per/offned showed tile mex 10 be a good /eoldrng mdrcator 01 growth m the Texas economy To compule the Index. t setected Ihe top 44 coun/fJes ranked bye)({)OIts from Texas rhese countrres accounted lor abovl 91 percent 01 rexas · exports m 1988  18  me:ISlI rcs lllm·etnents in real exchange mtes for 44 ("ountries, accounting fo r man.' than 91 percent of Texas t.'xP011S. To judge v.·hich Texas cxch<Jnge-ratt.· index W;).~ a heller leading indica tor. I L1sed criteria that arc very similar to thost, used by the Bureau or Economic Analysis (HEAl, u.s. Department of COllllllern..". in c h oo.~i ng \ ·;.ll"iabJc.~ fo r the nation;ll BEA composite index of leading economic indicators. The HEA scores \·;H"iahlcs on tht: basis of economic signifi (";lnce. statist ical adequacy. c}'d icll lim ing, husiness cyde C(ltlfOlll1ity. smoothness. timdiness .•mtl r(:'\·isions. The procedure I used places pa rticu la r emphasis o n business cycle confo nnit~· (sec Phillips 1988h). By usi ng tht:se criteria. the IlL·\\· Tcxas trade-\\'eighted val ue of the dollar was determi ned to he <J better lead ing Fede ral Rese rve Rank or OaUas  inciicalOr of mOH:'ll1en1:-. in the Texa. . . coi nci(knt index _! Con:>cquently. thl' new Texas dolla r index was substit uted fur the old measure.' Another change in till' TLJ was to reduce the weight gl\'en to the t\\·o e nergy \·a riahles. Recent resea rch has shown that Illc energy seclor is less imp0l1ant to the Tex;!s cconomy than il has been in thc past (for example . sec Fomhy and Hirschberg 1989). To account for this c hange. [ c()l1sidt."red dropping one of the two <:nergy \'a riables from thl' 'I'll. In l'\'alu<ltin£ wlJ(:thcr the real oil p rice or the number of drilling pcrmit applicatio ns was preferred . no dear answer e merged. Although the nu mber of well fX-'rnliLo; showe d stro nger lead ing abilities. its lead was shoneI'. a nd the series was mo rt' \'okllile th:m thl' rea l oil plicc. fiecause 1..':1('11 selies has equa l hu t separ..l1e ad vantages. 1 chose to weig ht eaeh series hy h;111" ils origi na l weig ht In effect . Ihis ste p comhi nes Ihc IwO sclies into o ne cnergy sector "ariahle . A fi nal e h:mge nmsid e red wa." thl: usc of the new experime nt.t l leading indl:x p roduct.xl hy Stock a nd Watson at the ~at i ona l Bun:au o f Econo m it- Res~arc h (NBEHl.' Altho ugh thc !'\BER le:ldi ng index tn(wes closet\' with thl: REA lead ing index. th ~ir constrllction is qU itl' diffe re nt. ,mel the ' BEl{ index co uld po.'i.'iibly :ldd further informalion aholll thc L".S. econOllly 10 the Tex:1.'i lead ing imlt;x. O n thl' bas is o f the tTilel;;t descrilx:d in Phi lli ps (1988h). Ihc ' fiEt{ kading index W:J~ s hown to be inf~rior to the BEA indl:x in iLo; :lhi lity to l e~ld mon,'menls in th~ Texa . . . coincident index. The ~ B I-:H index alsu showed no margi n:l l predkli \"(~ power on:r the BEA index. so the !'oJBEH index \,-as not included in the ne w Texas kac.ling index. The NTU \\,;IS com pa red w ilh tile TI.I. using the pron:du n_~ described in Phillip.'i ( 19H8h). Although re;] ks and troughs in tht· two indexes matched a lmost ex:!ctly . the b usiness (,}'de conformity n iteria showed tl1;.]( Ihc ~TLI had a stronger re];tlionshi p with the coincic.ll:nt index th an did the TLI. As shown in Chart I. the :-JTLJ mon: . . . dose to the old Ttl.' But during se\'eral periods. thl' t,,·o series di\'crge somewhat. In 1985. the origina l index was much "Taker Ix:causc of the larger weight gi\'cn to the energy indintlors. In late 19R'5. ho\\"c\·er. both indexes Ix:gall to plu nge as Economic R~iew - July 1990  Cha rt 1 Texas Leading Economic Indexes (January 1981 .. 100)  ,os  No.  95  go  85 1981  1982 1983 '911-4 1985 1986 1987 1988 1989  AS e)(p/amed m Phillips (1988a). I developed the Te)(as co· incident mde)( as a timely meaWfe of changes m Te)(lIs  me  eeonomy The cons/rucll()(l ot /he rnde~  IS  sllTlilar /0 lhal of  the U S COIncident rnde~ produCed by !he BEA The Texas cOIncidenllf)(je)(. however. is limited 10 IWO "~lfIab/es. while !he natl()()8l COIIICtd8ntIf)(jeX contams lour vaflables rh(> rwocomponcnls ot the rexas COIncident Index. nonagncul· lural employment and mduSlrliJ/ producllOn.  Ilave nallOllal  COImtCfparlS m the US COIncident mete)( like lhe ongmal dollar meaSUIe. Ihe new measure IS nol av8llable on a limely bas.s because of the lagged av8llabil. ily of Internallonal consumer prICe mde)( (CP!) data rhls  was nol a sl{}mllcanl problem because IIJI.' lead IIfTI£! of lho dollar Inde)( was much grealer than lhe load lime lor the other COl7JPC!l1fJfllS in /116 leading rnde)( The Oflglnal dOllar mdex was lagged Sl)( months so thaI ils lead lime would be more conSlslent With lhe other eompoilCflIS and to avoid prOblems With data availability roaccomplish lhe sameob· teetlva With the  new dotlar measure. I established a lag 01  IQUI rnOIl/hS • The N8ER raleases I/le /lew leading IIIde)( IllOflthly For m· lorma/1OIl on I/S conslruellOll. see Stock. and Walson (t989J , Tile new Texas leadmg mde)( 15 amplllude-adJUsted ThIS procadura sets the amplitude ollhe leadlllg index equal /0 thalollllecorfJCldentlllde .• TIlls procedure mak.es the com· pallSOll with /Ile comcident mrie)( more Vlsuallyappealrng. although il does flO( alleelllS predictIVe ebill/y In CharI I. Ille onglnal leadmg mde)( IS amplllude-ad,usled 10 lacllitale COIllJ)allSon Wilh the new Inriex  19  the largl' drop in oil pri<.'e..'.~ ;Ifft:ned aimosl e\'ery sector of the..' :;tate..' <,,"I,.·onolOy. :"/ot ie..·c ;11:;(1 th:1I Ihe \'TLI ~howe...'d mort: of an UP\\ ard tre..·nd on."r tht: past se..'\·e..-r,tl year.; than did the ori~in;11 index. Thi ... ,,!so is due mostly to Ihe wcightin~ of tht: e..'Ile..'rgr \·ari;,hk::., which \H:n.· !(t·n...r'.tlly morc IK'g;Jlin' than thl' re'>[ of [hl' indi(';lIor:. during thi~ pl'riod. Turning poi l1l ~ ill Ihe i\TLI han' had a sl rung rt·1:IIio n ...h i]'l wil h l urning point:. in [he Tex:l:; c.:oincident index. A!'o seL'n in Cha n 2, the :,\T l.1 Illfllcd d O\\ 11 fOllr m onth:; before I lll' August 19H1 pe:lk in the coincidt:nI index, and il reIXlllnded fi\'l' monlh:. hdOl'c the trough in .I\ larch  'l1,e leading in<it.'x then n.:lxllinded in .I lll}' 1986, l'ight months hdorc the ht'ginning of 11le..' statc':; cconornic ~co\·cl)·. Thl' inde..'x rose fairly stl'<I(hly until :kptcmhcr 19H- , wht'n il dcdim:d for Ih'e e.."Ullsecuth c month:; and then Ix:gan a pt.'riod marked hy shon sp:ln.... of SI ft!llgth and weakncs,... Tht' declint.' In the ~rru lalt' in 19tC Ill"r h:u'e lX:l'n signaling ;1 gt'O'i\'th slowdown. Since early 191:18. the coincidcnt indl'x h;'ls llu(,1u, "Il'd bet .....cen strength and wt::,knl's..... with only 11 ... Iight upward trend,  Using the NllJ to compute the probabilities of recession and expansion  19K3. The Jc..'ading inde..'x 1h ... n peaked in April 19th. 16 month.... hefon..' thc ('oim:ic!(.'nl indcx pcah'd in August 19f1'i. This lead limt" rlIa~' hl' c!<""C(.'idng. howc\'cr. IX.'C:llbt' 11ll' declinc in till' k;J<iing indl.'x was likdr signaling Ihe grO\\ Ih n..·o............ion Ih:tl 1X'!t:tn in the coinddenl index til lale 19tH. (For mort' dl'lails ahout $o(I"o\\,lh cyc1t':.. st.>c Box A.) fol lo\\ tIl~ a lonR and I~lI hl'r sharp dl'dinc of 11K' Il'ading index from April 19H4 until fNcemher 19H·1. till' index bt.·g;1Il ;I pat!t'rn of ~ain:; and d~­ dine:. with ;1 gr~ I dltal up\\·ar<.1 drift, '1111S St"'Crns ('nnsi.~tcn[ "'Ith the gl'neral patlt'rn of \\'~ak w()\\ til in the ("oinddt'1lI tIld!;!x . '11,t' .~ [et'p plttngl' in oi l pril'c:; hl'~inning in hlte 19/:1-;, ho\\'en~r. cau...;ed Iht' leading index to plunge and the Texa .... L'l'Onorny 10 declint· .~harplr.  Chart 2 Texas Composite Economic Indexes (Janu31y 1981 • 100)  n'  '" '" 95  90  20  In dCll'nnining the sUC(,'eSS of thL' Ic;Iding index in s i~n a l in~ urx:oming tlUnin~ points. using a fe;ll-lirnt' :lpprtladl I'> imponant. L()(lkin~ :11 P:lst (,bl:t to find pt.';Jk.~ and trough ... b c"s~', hut (lnt' IIlusl delennmc wht'f\ Iht' uscr of thc index could lx.' awart' thaI :1 turn had o<.:t'um,'d. One..' common rt'al"lnnc appm;Jch is th11l thret.' n)nS<.'('ul in· dt.... dint':' in till' leading indt'x si~n:11 an uJX'omin~ n,,·(:c.... sion and 'hft't' con:-<cuci\'c increasl.':' signal an expansion, Although thi:. proeec\ure..' lIa:; some \ alidity, Ilxent I\,.'s<.:,m.:h :;ho\\'.... th:l1. a! lea:;1 in Ihe caSC' of till' HEA 1L':lding index, :1 St'quential pmb;!bilil)' method ha:. <I hclI~r foft'otsling record (S('C Diclxlld and Rttdebusch 19R9). In :1 sl'qul'nli:ll prol"'lbilit)' :Iprnmch. tIl<,' p rohahil ity of an ulX'om ing I't'ccssion is calculated \\'IK'n lhe l'<"onomy i.~ in an t'xpansion. O nce :t rl'cession hq.:in.... thl.' ptnb:tbilil }' o f :tn expansion i:; c:tlt-ulat('d . .'\ prohabil it y of 90 1)I..'r(''(,.'nt or more is regarded a:; a ... tro ng si~ n a l . Expansion :md t'l'c·cs... ion c.::m lx' ddl nccl in IL'I'Ill.... of chan).te:-. in lcl cis or gl'Owth ratl.'.... Thc st'quL'ntial pmhabili1}' tnL'thod ".~ de\ d 0fX'd hy i'\eftd ( 19H2J liSt'.... 1\\ 0 principal Slt'p... to dl'termine Ihe proh:lhilit}' of;m up<:onling I't.'ces:..ion or expansion. The fir:"1 step i... <.k-Icnnining Ihe likelihood Ihal Iht' current ch:m).tt' in the leo\ding inde,,- would o(xur during a husine...... ll'clt' e..'xpan:;ion or t'ontr:tcliotl . Thi .... b c.:akulated hy tookin).: ;It p" ..t dm:t [0 set· ho\\ oflen ... imil" r t'h:lI1gl.'.... took pl;1("t, in exp:tn:.ion... :mcl eOnlr:Klion... , For e..'xampk-. if tht' iL"ading index incre;Ise:. I per<.'e..'nl. :tnd in tht' pasl Ihi ... ()t'<" um:d 1') lime... w h ile :-.ignaling exp;lIlsion:; and only once \\'hik'  Ft.'tIcral Rt:lK'rvc Bank of DaUas  BoxA Growth Cycles Analysts measuring the business cycle in Japan and many European countries use growth cycles more often than the classical business cycle. A classical business cycle is marked by periods of growth and decline in the levels of overall economic activity. In a growth cycle context, however, a recession (expansion) occurs when growth in the economy declines below (increases above) its long-run trend rate . The Center for International Business Cycle Research (CIBCR) at Columbia Uni· versity studies U.S. growth cycles. The CIBCR analyzes changes in many series it classifies as coincidenllO the business cycle. In each series, a growth cycle turning point occurs when growth in the series goes below or above its long-run trend rale. The process of establishing the longrun trends is not straightforward . In computing Ihese trends, the CIBCR attempts to measure long-run rates of growth that are allowed to move over time but are independent of shorter cyclical movements. Using a computer program to aid in the selection of these turning points but still allowing subjective decision making, the CIBeR has established agrowth cycle chronologyforthe United Slates. To define growth cycles in the Texas economy during the 1980s, I took a different approach than that of the CIBCR. Because  ~ign:.lin~ comrn(.1ion"'. then the pmhahilitr is hj~h that the current change i~ ~ignaling ;10 t'xp;m.sion . The next step is lIsing the previolls pcri(xl's probahility to ...trcngthen or \v(.';lkt'n th(,' pmbahility c•• kulaK'" for thl." t'llrrt'nt pcri(xl. For eX;l1nple, Ir the leading index h:.d been declining for many months ;Ind then jumped 1 rcrcent. it may :It fir.;t be difficuh to distin~ujsh if tht: jump is a Il!mpo-  EconomJc Review-Jul y 1990  the Texas economy generally experienced strong growth during the 19705 and weak growth during the 1980s, it was difficult to estimate an underlying long-run rate of growth over this period . Instead, in the case of growth recessions, I looked for periods in which growth in the Texas economy slowed significantly and was close to or less than zero. In the case of growth cycle expansions, I looked for periods in which growth increased significantly after a growth recession. I first examined movements in the Texas coincident index. A plot of this index revealed two distinct periods in which growth in the economy slowed but remained positive. The first period began around the fourth quarter of 1984 and ended with the start of the classical business cycle recession in August 1985. The second period began in early 1988 and continued until the end of the data in November 1989. The plot revealed no growth cycle expansions other than the normal classical business cycle expansions. Statistical tests provided further evidence that growth cycle recessions occurred in the two periods of weak growth. In each period, the average growth rale in the coinci· dent index was not statistically different from zero at the 5-percent level of significance. The average growth rates were also statisti· cally different from growth in the earlier stages of the expansions.  r.uy hlip or a true signal. If the jump is followed hy another I-pcrc~nt increase, howe\,cr, then the prohahililY of cxp;tnsion should incrca.~ .1> Chart..., 3 and 4 show Ih~ prohahililies of rL'CL'Ssion :lOd expansion for T~xas from January 1981 to No\'cmht'r 19R9. As shown, the prohahil· ity of recc....~ion rose alxlVe 90 pcr('cnt three months bdor(.' tht: pc:lk in September 19A1. Zl  Chat13  Probability of Recession in Texas Petoent  GR  R R  r  R  GR  --__  InconSlruc/1I'I9 lheprobabiltry-<ll·receSSlOnIl'tdex lot TCI(3S I utiN:esomeedJu$tmenls tolhe Nelrome/hod fr~DreboId 9Ild RlJdebusch( 1989)  The 1..51 OOpSimern IS due 10 /hefr  68fflIIf hndtngs fllal 1116 ptobabUlry 01 recesSlOfl IS not  a  /tJoC/JOtl 01 lhe /engIh oIlhe CUftenl rflCO\lfHY The SI!O(1rId sa,vstmen/ 1$ 10 ltdopt /he/( use oI/he f/OfmaJ detISIIy /tJoC. IICII. IfISlf!ad 01 detNlflO IheprOOllbihly dlSlribu,lOtls drectly from h<51orlC8l diJta The Illrd adJUsiment address6s lhe tSSVO lIIal II ,/Ie probaO<Jlly 01 rcccSSlOtl or cxpanslOtl reiJCheS I. /hefl BII foI/owIng /Noba/JIbl.es will also equal , To allow more f/en/JIlily .,., lhe CqUatlOtl tIfI(j s/IN enable the ptlOf orobal)lilly 10 affecl the currenl probability Sl(JnlfI. ca,,'/Y. I Sal an ufJPt!f bound of 0 95 Oil the or"" proba/JIMy lhall/JfJds mlo /ne recursive formula A drawoocl! 018 kJadmQ 1Ildc)( /nal Signals growth r(]Ces· l/lal. once a growIII receSSIOt7 has begun. /he leadll1g lI1d/Jx may be of '11IIc use m determIning whethet a classICal OuS/Il6SS cycle rccesslOt7 !"IIlIlo/low In applymg lHe probablll/y-o/-I9CCSSIOfIIO£fnula 10 lJfow/h cycles ,n the United Stales and 0Ihe1 natJ(]()s NlCfflNa (1900) conr:en· Itates solely on the f}lowth cycle and does flO{ lKkkess the SIOi!S 15  queSl/Oflol Shlfllllg 110m slOwgrowlh todecllne ThepredK;·  110110/ svch a slltfr may ~ above the bounds 01 me Jcadmg  a/rhough II deserves somo;o research the f>I.Jt7Ibcr 01 paSI observa/JOiIS IS so  1tIdIca10l IJppfO(tr;h  CerlalN;. IfI Toxas  smaIIlhat lillie csn be Ieamed atxJul /hIS aspecl 01 busilJeSs  ",.,  DreboId 8fId Rvdebusch (1988) WX1tesscd trltS Issue In evalua/1fIg the petlomlanr;e oIlhe  BfA /tldex oIleadHlt}  BCor/O'fJOC 1fIdK:1I101S lhe8Ufhcxs fooodwbsJanuallyweaker  oe<fofmance when USIflQ/ho rCiJ/·1Jme /f>dex values • The /Odex cs/CvIa/ed /hIS way dl/lelS sllghlly fr~ lhe NTL/  The calcula/1Ot1 u$6d /Nt 0l'lgm81 TUSS valve cllhe doI/ar  22  Following th t.> 1981 signoll and Ihe stan of the suhsequent rt"cession , the prohahililY of expansIon increased to a rate h iRile( than 90 percenl in Feb.. ruary 1983. 1\\0 months hdore the recon~r)" in the coi ncid'mt index . The prohability of recession Ihen rose aho\"t~ 90 pcrccnl in September 1984. ele\"t'n months before the peak in the coincidenl index and ;rbout tht, samt' time or sligbt1r before Iht.> <Jpparent growth recession began . The proba· hi l ity of exp.msion then sign;rk'(l an upcoming ex· pansion in .I:1l1U<1I1 19H7, Ihn:e months before Ihe exp;ll1sion actll:l ll y began. The ne."\t siRn:l1 from Ihe index ClIme;: in Octoher 19ff'. when Ihl..' prohahility of recession roS("" abo\·t.> 90 percent. This sign;11 (;;Ime ht·fore the grx)\\t h ,.el"e~<;ion Ihm hegan in c;lrly 1988. Statting in c;ll"ly 19M. the prohabililY of:1 growth cycle cxp;msion was c:'lkulalcd - The probahility of a growth cycle (.'xpansion nuctuatw. hut in onJ)" one monlh did th(..' probahililY reach higher than 90 pcr<.X"nL Thi ... signal oCClllred in ,.Iay 19R9. and it is too l:;lrly to jmlg(.' whether a l!lrowlh expan· sion follow(..'(1. O'·eral l. changes in the le;lding index apJk'ar to it:;ld dl:mgc:-. in 1hl' coinciden t imJt:x . The proh,lhilitr o f )"ecc ....sion reached 90 pcn:cnt before (:'n:'l"Y re<.:c:: ......;ion and gro\\ Ih crcle rcces....;ion in this limited time ~ r iod. Thc Icad time. how(..'\·cr, has Ixen relath·d} short. about Ihrtt month!'>.  Sensitivity of the NTU to revis ions in compone nt data The ,·;dlle of Ihe: ~ TJ.J in :'lIly gin:n rnolllh is l"L'ds(::d as the: cornponelll data are re,·ised . An l~'·alll:lt ion of lite predict!' t' content of the index Ihal i...; b;lscd on the nn~d n.'\"iscd series, such ;l<; the e\·"lualio[1 in the prt'dous section. could I~ qllile different from one hascd on the first estimale of Illl' index ." From Septemher 1988 until .\"m·cmhcr 1989. Iht"" u riRina I "1"1.1 was proc.llI<:(..'(1 on :1 monthl y basis. ;mel Ihe dala were slo["{.'d. By llsing thes(;' data. it is pnssibk- 10 conslru<.·, the :\,TLI on ,I re;ll-time hasis 10 endll;Hc ho\\" n..'\·i!'>i{)n~ in Ihl;! index \\ ()uld h;l\·c alTL'Cled il!'> pcrfonnann.'." E, e[")" month. the uaW ror the pre'·iOlI!'> !'>e\·en momh!'> of the t\TLJ are 1"C,·ised 10 incorpo· I";.lte !"C,·isions in component dala For e:lCh 1110mh  Federal Rl..""Scrve Bank or OaUas  lISC thh mfunn:tlion to impro\'(: the pn:li1l1m:uy estimate:.  Char14  Probability of Expansion in Texas  Summary E.e~  1988 1989  fmlll .'!I..'plellli>c..'r 19M lhl'ou~h ~I .. ~, 198<). [ renmlcd tlK' l1r..t to the ~·\(.·!1th estimate of the ('hang(' in till' index, U;!sl.:d on thi~ ~;llllple, thl' "tandard de\bllun of tlK' 11.·\'i.... I0!1S from thl" fir..t to till' nn;!l e ..tilll:lle \\ :1'" 0,25, Thi ... rcprcsenL..; a 1II1xk-r,lIe dq..tn..'t' of R'\'isioll For example. if the prclltlllll;lfY l'stim:lIe sl!m\'l'd :1 dwn}.:t:' of 1.31 rx:rn'nl (lj2 \\:1 ... the :1\l·r:.I).!:l· absolute \'alue of Ihe IX'Kl'ntage ch:tngcs during thi.... period), then Ulll' n)llld lx' 9O'rx'I'('l'llt ('(JIlfidem that Ihe final l·:·ail1l'ltl' nf the d langc \\mild 1')0.' lx-twcer1 0.9 1 IX'rc('nt and I. -,'S ]X'fcent. a. . sllmillj.l normality. TIl1:- perforrn:ll1t'e l·omparl· .... f:l\·ol~lbly \"ith that of the BEA n:ltional k:lding index_ wbk.:h gener:llly ha:-, had 111lJ(.:h i;Jrgl'r re\ i... ioll .... I" \\·hih.: tilt' I'l'\'isinns in tlK' :\"TLJ l1:rn-' hl--en lllc)(.icraK· 0\"'·1~ 1 1 1 . Illl' c;lrlief rl·\'i.... iclllS ~e..:.H.'I~III)' 11:1\e b...·...·n lar).,:ef lh:m bter fe..:\·ision ... The aH.'r:age..: allMllute ,;II\lt, of till' re\·I:-.i(111 in I he ~rct'ntagt" (·hangc ... in the ~l \\,;1" 0.1:; for the first R'\'ision, 0, 1- for the ""''i.'oml. 0.(,1() for till' third. O,OR for the fourth. 006 for th ...· fiflh, and 0.02 tor the sixth. AU(>lIK-j' 1l11pol1:mt ;"'IX'l'l ot thl' prl'iiminary e!'>timale!'o, othn th;m how dn.'ot;' the)' <Ire to the..: tln:.1 \:lll1es. '" \\ IWIllel' tlll'Y :.r...· dlkicnt in:l s1:1li .. ti(::1i o;cn:-;<:. In the limitl'd .. amrk· ~ricxl. the prdillllnary l·!'>timatl·:-' of Ihe ...-1l:.llgl'S in the i\'TU \\,,,,'.'l' lInhia'o(..·d and l'O'icil'n! e ... lilllator.. of tlK' final '·allle.... of till' '\'TI.I Box II dt.'''l'I'iI>t:~ th"" te~{<; for thc."t.· pro'x'11H.:..... II lhe pR'liminal)' estimate... "'ert· h.;J.... t·d or indlkio..·nl, th .... n the l'l'seardll'r cou ld F.COflo mjc K("Ylcw- July 1990  Allhough Illy ori~inal ('ompositt' index of Tl'X:I .. k;lding e('ollomi(' imJiCalor.. b scnsiti\'e \() chanw.~ .. in the Texas economy. I uti lized recent information 10 nm.~lnlct :1 nc'\\' Tt'xa:-. leadi ng inde:.: (\,TU ). T he :\"TLl tI~ICk:-. thl' old index dosdy hUI shou ld predict futurl' turning points in tht;' Texa.~ economy better than the old index wou ld. III c\':lluati ng the !\,TI.I \\'i1h the st'qul'ntial prohah ility method. I find Ih:1I the new inell':': ha .. 1X'lfonm::d well in pn:dkling growth cyell' tllrning points. AI'l(I. during a short eX]X'linK'ntal lX'liod. tiK' rt::\·jo;ion:. JIl the :\''1'1.1 were sm;.11. ;111d the prc..:liminan e:-.tinl<lIc" Wl'!'t.' \lnhi:I ....."(1 ;mcl l·mden\. The results of thi ... sludy Impl)' that tht· i\TLI i... a u..efultool III e\'aluating tiK' ('h:mging condi· tion.. of the Tex;ls economy. l'sed along with otht'r loob. !'ouch ;l!,> fUrt'Glsting modcl.... clt:mo.. Rr.lphi( and indu ..try .~lu(Ii~'i_ and judgment:!l ;malysi..... the "1'1.1 olTer.. the ;maly:.! an oppor, tllllity 10 illlprO\ e hIS fOI't,.'(':I."ts of tlK' st:ltt' l·(·CIIl(IIllY·  because of  ~  d,lIlCuIty If!II'OIWKJ  If}  ,econS/fUC/lOg paSi  valtJes of /Ile flew r6Xas value 01 me dollar on a rw/·11tTIf! b6$/$ P;JSl CPr eta/a lor the 44 CCtJf1lfl6S III the ne-,,, Texas value oIlfIe rJoI/8f ''l8fe not aV8i1a~ 81a rC8SCtl8b1e COSl  on a reltl-lme ba$lS W"", the SlJmpIe peI>Od  I US8d lhe  ang.naI Texas """" oflhe dollar b6cavse ,1 had already t:/eI9fl CMCuIIJIOO on 8 feal-/ITl(! bils<s and SlOt00 BoUt rndcJtcs cornaon many 0I1he samIt CCtJf1/fllJS however and (1!MSIOf1S on /tie Ot'f1lll8l doIliIt rndfcafOt sho.*1 serve as a f}OOI1lfldtcRfOt 01 f~ If! /he new dollar rnd/CiJlOt R8sui/s /rom DtebokI and Rvdebusch ( 1968) SIlOw /Ilal lor Oecembc< 1968IOJalllJ8fy .981. oncCOCJkJ09 9tJ.per-cenI cOtII/deIlr lflal "the prelmHlaryeslrmale oflhe BfA leadroO It'tde~ IflCfeased 132 percent IfIen lhe Irna! serres change between 0 09 perC6fl1 artt:1 2 13 percenr  ~'IOUId  Box B Statistical Properties 01 the NTLI Preliminary Estimates The tests I used to evaluate the statistical properties of the preliminary estimates involved the following ordinary Jeast squares regression :  Final, - A + B Prelim, + B" where Final, is the final estimate of the percentage change in the NTL! for period t, A is an intercept parameter, B is the slope parameter, Prelim, is the preliminary value of the change in the NTL! at time t, and 8, is the error term. Six regressions were run , one for each of the six preliminary estimates . Autocorrelation functions derived for the residuals of each regression dampened quickly, implying stationarity in the errors.  This was not surprising because the variables were measured as percentage changes. Box Qstatistics of the residuals also indicated that the residuals were white noise. Because the error terms appear to be stationary, white noise processes, it is appropriate to use the standard errors on the intercept and slope coefficients to make inferences about the value of these coefficients. Likewise, the use of conventional F statistics is appropriate for joint hypothesis tests. In all six regressions, the joint hypothesis that A was equal to 0 and 8 was equal to 1 could not be rejected at the 20-percent level of significance . This result, along with the error term results, implies that the preliminary estimates were unbiased. efficient estimators of the final value.  Federal RHen'e Bank of OaJbs  Re feren ces Diebo ld, fr.lOcis X .. ;md Glenn D Ruddm:;ch  (19M), ~ Ex Ante Turning Poin! fCJr(,'C.Jsting with tlK' Composite Leading lockx : Financ..and Ec onomics  Di:>\.' lI~sion  St:ne!<o, no, iO  D ,C. : Ho;mJ of GOH.'rnor:-. of th~ Feder.ll Ik .. ~I'\'c Systc:.''ill, Di\'I:-;lon of Rc<.;ean:h ;lOd Slali~tics, October)  EC(JIwmic hulicalOrs  ,\(,/1' Approtlcbes  lIlId  FrJlt,'ClI,~Ii"R Hecord(,  l-o, K:lj;d Llhiri and Gc..'offre}, H , :\Inure (C.llnhridgl' : Camhridgt" lInh'cr:o;itr Prcs~, forthcoming),  IWa~hington,  _ __ , and - - - (1989), "Scoring the Ll:ading Indicators," JOIIl'llal qf I1l1silU!S.I' 62 !July J:  l'hillip:-., Kl'ith It (]<JHHa), "\'e\\ Toob for An"lyzing tlw Tex;!.') F.nJnomr: Indexc!<o of Coincident and Le;lding [(:onoillic Indit':Hors," Fedc!~l l Rt'M::n'e H:tnk ul D:dlas hi.'Vl1umic /(el'iell',  III I\', I - B,  369- 9 1, romby. ThOI11:IS B .. and Jo.~er h G, Hirschherg ( 19R9), 'T exas in Tr,lIlsitinn: Dt"pcndcnn' on Oil ;1I1d the t\a tion;l l Etonoll1Y: Fedcr.ll Rl' SC1YC B~lnk of Dnll;ls E c m/OIJI/( Hel'jell'. )anu;uy. 11 - 28, ~dtci,  S;dih S , (19HZ), fi Optinl.l1 Prediction of Cyclical I)ownturns," ./rJllrllfl/ q/ £COllOmic 01'IU/mics lllld Cul/lroll (:"\m'cmbc.:r): 2Z'S-11.  _ _ _ (19HMb), "Thl' i)C\'t'i0plllcnt and l'.,>c:o; of Regional Imkxcs of Lc<ltl i n~ E('onolllil' Indicator,, : Federa l Rc:-.el"\'e Ban k of i):tll;ls Rt,:o;cardl P:lrX'1' no, HHOH (Daltls, '\'oYt:,mlx'rl. Sto(: k,jamc:o; H, ~ll1d :\l:lrk \\ ' \\'""on 119H<)l, " ll1dl'xc,~ of Coim'idenl :In<l I,l'ading Eo ... nomil' Indicator..: in ,\BEN ,1JtlCltJ('C.-cm(lmic.~ .'llIIlIIull9H9, ed , Olh'icl' Jc;m BJ:mchard ;1Ilt! Stanler FI:>\.'hcr (c.uuhnd).:t·: ;\11'1' Prc.::-.s):  3:;1 - 94 , :-.Iiemira, I\ l ichad p, (1990), ~An Intcrnational Ap' plk;llion of :"\dki's I'rooahilitr Approach for Sign:llhr)( Growth Ikct!ssinns ;md Rl'CD\'t:Tit':o; tJsing Turnmg I'oinl Ind ic;ums: in U'f1djll~  F.OOJlo mle Re view - July 1990  Tcxa:-. I'k'p:lrtml'nt of ('..ol\llllc.:rct.', Rc;carch and Plannin~ J)i\ i~i()n ( 19H<)', fJip.bliJ!.bfS (8 1987 (llId 1988 Texas EXjJol1S (Au:o;tin, Ot,tohc:.'o,  "