*) '~ o ) .9 NUMIBER 35 *a F Pecisr Operations Evaluation Department October 1992 Integrated Urban Projects: Experience in Ecuador, India, and Brazil Taking a "sTIeMT.bistic` ap)proach to operations ifieldssonei practical recom- viation and had broadly similar ientropolitall dtevelopenitet /ills strong nenldlationIs for- planimiwiig unid im ple- goals and1 approaches. appeal. Ans til' Bank expanded itS urban metntini, ciirret't projects. *eflhliil,> ill tilt Ial te 19,70s adi eatl - if The Guai7t11aquil Urban Develop- 19S0)s, it iiundlertootk a series of inteyratel ost of the Bank's urban meat project, the first of the four to iopcratiolIs ill hop)es thlat thtese Coulld I / development lending be approved1 (see table), was the first have a greaiter impact o01 uirlban lininS approved in the 1970s aned project for Ecuador's largest city Conditionis and incomects tiuiii separate early 1980s was for low-inicome (population then I million), and the projtets ill sulkSICto0S SuIC/I as h1oulSilln, housinig or urban transport projects, first ulrban project in that country. smliall industrV, or transport. * but rouglhly a third (or a fourth of Designed to iniitiate large-scale pro- the nuimber of projects) was for grams of hlousinig development and Experiencc has shozwn tlat tlic multisectoral, or integrated, projects. job creation to beniefit low-income adlv)ililtiiges of initegratediI projects inc people, it contiained c*omponents for Oen/, hlardl to capture. Rowent OED Integrated Lirban projects view sites and services and sluIM upgrad- aIIidits in three countries ihliistrtite thei urban andi regional problems as ing, support for smnall-scale enter- lliit7tatiOilS, p7rolellMS, andll risk;s associ- those of a single systen. Compared prises, anid technical assistance. As atedl zith integrated urbin proec-ts. WithIl most urban projects sutpported later reformula1ted, it also helped to SIuc/I projectS ts 1epcild ea'ilil o01 aI Slilp- by the Bank, these are complex rehabilitate tlood-damagedL urbain portiove niacrioccnionomic, policIi, inistiti - projects withi ambitious institutional infrastructUre. tiolial, and local politicai ctivironimieult. goals. Typically, thev contain com- TheYr have rarelY occouiplishel theiir ponents for shelter, infrastructure, The project's ma in objective, goals of institutioinuual atihi di'l' miii trainsport, solid waste management, however, was instituitioral I and better filliaticial nillanielit in agencies bUsiness support, helalth, nuLtrition, provuidinx urlani sen' ices. Tici reqluire aind edLucaltioll. Whel introduced, tin;i Coini,iitnie,it 1) local 11tiionicws a7s in tegra ted projects were tholughit WC'l (as; striliSn1 n, c) (isi ilt/and in ilti- likelv to Iiave a greater iipact oil *See Performance Atdit Reports ficeteil Bank inu;zoleeoiet,l incluldiig urban dlevelopimient than discrete for "Ecuador: Gnayaquil Urban Le! intsiec si/le rz'isii1ll. l<(( (copi iz- prlojects witil a subsectoral focus. Development Project", Report iths til fficalttii,,, the liVaik now' Costs of infrastructUre and transaic- No. 10238, Decemnber31, 1991; I,Oqil/ lirccts Illost of its Ilrbium ahllin tions wvere expected to be lovwer, and "IIndia: Secotnd Madras and to) immla'ii ni(tiir silbSe(tO/S npleen taotion d elays sliorter, in Kanpiir Urbacn Development tiese projects than in subsectoral Prp`ects", Report No, 10579, 9 Interi'iitioiis wor/k beltcr i f ifti ;i i ne projects. Aj!ril 22, 1992;aid "Brazil: ctlecti/ce iltt clscl/ focuIsetd, bult theiwl Recife Metrapolitan Region De- iictil 1tbechose;! mt pll nllltlml)lin I/ith Project goals velopment Project", Report No. iilht otif on vrall pIlaint Il1{- urb1an man- 10832, Juniie 26, 1992. OED i'l'It' Multn iili(n einlts. Dcaiih,uu; The tour projects audited were reports are availablefrom thte In- Zl'iti thli' inthrfiice aCmn0i,' tle t dicrte tvpical of integrated projects in ternal Documents Unit andfrom ilterve,itiimis is -till n maior chml cn!i)1. Single large cities oir metropolitan Regional Intfonntation Services Vhe re:'iiw ofecvpericncc xith hitt-.ratedt areas. All focu'-,ed on po verty alfc- Ceniters, I tration, anrd financial management Project Facts associated wVith urban andn metro- politan region development. Pr ! , 17r]ttct . ,-11'roEt: Close4 Lai,n.Alwwvit Ech lof the projects relied on I - u,Uqio,'" severil differenteXecuLtilln,'aenicies __________________ _____-______!_____ and h-ad complex implementation Guavaqud L'rban Deeelopnient I;I 31 arrange-Ients involving at least two C GWIdac!.Jil Lfrban D)evelopFnient IU7-9 10Y 1, * i b ~~~~~~~~~~~~~~~~~~~~~~~~~~~levelso>;:7novernmenlt In Recife, the extremecase, the project involved adras Second Urban 1980 1988 42ctiities and 33 federal, state, 7 Development and Ic.CJ agencies, including a Kanpur Urban Development 1981 1987 25 dozeniSeparatemmunicipalities. In RIcife MItropolitan 1982 1989h124 \'lad ras,Kanpur, and Recife, overall S Recife Metropolitan 1982 1989 124 project coordination was the responi- Region Development sibili tv uf a metropolitan develop- -_____ !_________________ i _______________-__________ ------ mentaulliority. In GUayaquil thlis role s-va sclivided between the mu_- nicipodi tv and two central govern- nent hCLusing agencies. policy development. This WdS to The Kanpur Ulrbanz Derelopmenitt Implerlle ntation, results be promoted in the coLurse of project was designed to help India's expanding7 credit anid teclnical eighth largest city (1.7 million All Otur projects experienced assistance to small enterprises, inhabitants in the early 1 'Nl. to consi dercible delays (on average 1.8 strengthening of municipal efforts to increase the supply of servicecd land vearsi)ai:id otlher implemelntatioll serve low-iniconle areas, andl build- to the poor by providing livv-cost difficulties. Implementation prob- iilg tip the experience of the national serviced plots for housinig and small lemns wcre most serious in the two 1hOuLsing bank in financillg shelter buLsinesses, and to address tlhi worst latirn .nerican projects. In both for low-income people. environmental condlitionis in pri- cases, ocal politics ald institltional vately-owned slun areas in atid rivalries, whose importance hladi The two projects in Inidia arounld the citv's core. Other goals been disvc4ulntecd d Uring appraisal, consisted miainly of shelter improve- were to strengthen key urban playcol a mrajor role, as dLid deterio- ments buLt also contain ed comp- instittitions, develop a stritegy for ra tin~\ivcroeconomic conLditionis, onents for urblanl transport (Madras), state-wide urban developnient, and whici lad versely affected couniter- traffic management (Kanpur), basic obtain full cost recovery from the part futnd in4g sanitation (KanpUr), solid waste, project's shelter compollents. municipal mainctenanlice, iiind r n Lre !hit/ ( 1thll1ii project, imlpe- institutional strengthening (botl). The Rccitc Aletropolitat gion mentdtiioL of almost all the compo- project was the most ambi tious of ncrnts u ncder municipal respotnsibil- 'Tlhe Secod Madrnas Urban the four. Recife is Brazil's fiftlh larg- ity providl problemnatic. Ihe nui- Dczvlopmoit'n[ project was dlesignled to est city (populiationi 3 million ini nicipiiii tp v never reachedi agreement help fndi'is foulrth la rgest city I980). This innovative project cotn- on CrO'st recovery with the beneficij- (populLtion tIhell 5 millioll) to fur- tained comllponent,ts to: rins ) t he upgrading projects. ther develop and expand the low- FLhunt in J ,andl conltract execu.tion Cost solutions to urban problems * improve housing, turbon inlra- prOb lenis for a sites and services beguln under its predecessor. ft strLCtLre, and services within tthe sc herme COLscLd lon11i inlpl1en ltlt.Itia t suipported HtheconLtinued re-orienta- existing urban fabric; d'iavm, promptillng nei,ghborin- tion of shelter and infrastructure faniliest) invade the plots and inv'estmenlts in thle MaIddras Nletro- * provide infrastruCtUre and lead.li lag hIe municipllity to 'i ,Indon politan Area, to miake tlhem more services to support, compleme nt, the uv,orks. respon-sive to the needs of the] Urban and indtiLce desired pattertns of met- poor. It sought to stren,gthe the ropolitan develOpm11enlt; TI o.1.,Il. of thIe sn.u->c le i nstitu tions responsible for slhelter, enterpr iso' credit pgramlll proved infrastructure, and trainsport, to * enhance inicomiegeneration, too con-iplicated. Rediseon iit ralte encourage grea ter emiilphasis on cost especiallv for thle urban poor; and set bx detlli central bank were t0) lox\, recovery, a nid to Support effective d1isceu-Ptli1g thIe partie ii ption ot metropolitan plaLnning and capital * strengthnthe institutions re- priedte interniedirro. \s.aresult. programm ing and bUdgetigl,. sF ponsiblc for the plan ning, adrninis- b.tll 0IC-th L;lln uLpgYLrdil' Jnd l I OcYtobc'r I Q-Q9 I Ii businiess components were event-- 1 ally dropped. The project was Compansons from Brazil and Ecuador retormulated, reducing the municipality's role, expanding the OEDevaluated the Guavaquil The \L-rt!teazt Urhl;rn Flood RreC"L- serviced sites component executed and Recife projects JoLdtlv with other 5tnidtwon project i 1985J was an "emer-l bv tile national 11ousilln institutions, urban projects in Ecuadcorand genc " program to help rebuild and1 addSing> H1oodi reconstruction Brazil These cluster audits permit- housing and infrasrnucture dcrlaged i works. Even so, it achieved few of ted an assessment of the etfecti, e- by severe flooding in Brazil's poorest its policv, institutional, or financial ness of "sectoral versus integratEd region. Though start-up delays led goals, inclu1in: the strengthening of approaches to urban development in project resources to be reallocated muLniCipal finances and service de- the same countries over roughly the from housing to infrastructure, the liverv and cost recoverv from final same periods. Operations of the project successfully installed storm beneficiaries. i former type were more straightfor- drainage and other flood control ward to implement and they gener- works thrmughout the Northeast, fn Recift, miacroeconomnic insta- ally performed much better. affording protection to many low- bilitv, shortages of colunterpart E -cuador'sVatioial LnL-inct.rn, income urban neighborhoods fulit-is, and political-atlmiii-iiistriltive I Eudrs.,toll Ael.trl hinds, andtupol"Ital chadnies t A hree Fiviising project (1482) iollo-ed di- The Parnrna M4arket TLot. (" lprove- and institUtioall changes atm all utcerefulthousin levels of governm ent a dversely rectl n om the s uc u housi ng . project I 0S , i icOmpoilent ot'the CtiaYaquil opera- involveda combination of loans and~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~cmbntinoflan n affectedl project implemneiitation. lition Carried out in 15 large and grants to loc.al governments. It too Being botil multi-institutional and medium-sized cities, it supported1 a - surpassed nicist of its ph%sial targets place-specific, thie project proved shift in national houmLng polics to- - a nd achieved most ot its insuhtithonal . vulnerable to administrative wvard provision ol aftordable sheller and financial goals. Among the rea- changesand political rivalries. Dur- for low-income families. Implemen- sons for success: ing the implemenitation periodi tation was highly satistackir due to - Recife had a succession of six gover- the projectes comparatively simple * The project was based on an exist- nors antd at least five miayors, as well design and good perfomiance by the ing local initiatj' e that *% as .ble to as seven Superintendents of the executing agency. Physical targets maintain strong political support project coord inatitng agencv; lLt.mer- for both low-cost housing and home even when control of the state gov- OLIs clhlllges also took place amongr improvement loans were exceeded. emnent shifted from one party to the mayors of other metropolitan I The project suus4ullx introduced another. m1unicipalities, state and local policy changesin housing fincance * Most of the subprojects were plainiiiig ai se tedtnh d participating small, with relatively simple designs ann lcadis of project execrtitig ages, - inistitutions, that facilitated implementation. a n d h i a d o f p r o j c t e x e c u i n g a g e n * T h e p r o j e c t w a s a b l e t o m e t h eies. Each transition, esp7ecially Brazil's Tliird Urban Transport priority demanids of a large number when dlferent poltlcat l pa1rtes were project (1.981) likewise exceeded its of towns, providing it with addi- involved, broutghtatleast some physical targets: more intermediate tional political support. changes in personnel. procedlures, cities than planned benefitted from * Because it covered several towns, and/or priorities, affecting the pace local transport and traffic manage- risks of delayed execution or non- and. soimetinmes the direction of iment improvements, and low-in- exectition were spread over a large project exCeCutionl, I come neighborhoods in more urban number of subborrowers. areas than planned benefitted from * The project was coordinated by a The project was revised twvice the paving of bus routes. The competent and motivated team in a and several of its original compo- project also enhanced the opera- state with good public administration nenits wvere eitlher never uL ndertakenr tional efficiency of the national ur- and comparatively strong local insti- or substantiallIv altered during ban transport agency. tutions. implementation. Othier compolnelLts i- wvere onlv partly execuLted or proved to he of little lalStillng eConlomliC valule, innovative and ambitious design, requiring periodic tariff increases bv ev tilougl Somlle of the infrastruC- dlifficult to realize in the best of cir- thie citv bLus COm1p1anJLy1 illitial illStitl- ture investments antd 1malY of the cumstancies, couLdII not witlhstanvd the tional problems with the sanitation slum inmprovemnwt works resultedj econiomiiic, political, and institutiollal component of the Kanipur project; in tanlgible bene-fits to low-inco me turmiioil of thle implemenltation period,l alldL land acquisition difficulties in people. 'I'lTe project's financial anitdL botlh operations. institUtiolnl aL ccomplishmnents wvere In the two 11j0nh cases, inmple- negligible. inlentation proceeded more smnoothlv. In the end, the Madras project Exceptionls Were a suspent1ion of dis- substantially excLeLee its physical ULltima,telv, the project was the b1ursements in \aSidras, because of tar-ets tor sites and services, .slum vic-tim of its, own p()11 l\itv. Its non1-coMpliance wvith a covenant up Igradin , a d bs procur en t ()ED Prc;Cis i i i i Operation of the bus comiipanyv experienced task manager during Recommendations improved considerably. In both most of its lire cvcle This helped to cities, shelter investments have ease tensions witlh the borrower * hI urban tvtdcs;&7 pl1 ra7cal, noticeablv improvedi the environ- when disbursements were sus- anld stcu'l.'isio,i. panltclosl t menit in thle neighborhoods affected. pended. Similarly, the Kanpur political flictors. Borrower coirty.t- Both projects hiad1 acceptable project was supervised by the same ment to project goals anL i viabClŽ re-estimated rates of return, con- core team that oversaw its prepara- implementation set-Lp are ke . firmin- the significant potential tion and appraisal. Both Latin reqluirements. Politicalt tcturs wvi 11 economic impact of developing lantd Americani operations, in contrast, always affect project perfomniana2, and low-cost 11ousing in large Indian experiencedi several changes in task blut their potential adverseini ptt cities, where the supply of both is managers, while supervisioni mis- can be reduced either by tlking ofteni severelv constrailned. sions often had insufficietit personinel account of such factors irv tailoril- to assess performance adequately or project implementation p-eriods,;r-r But even these projects had provide the assistance neededi for all by designiing operations FlexibL[, notable shortcomings. Occupation project components. enough to accommiiodazte shifts in of sites and services areas in both priorities whein administratiors cities has been slow and manv plots Lessons change. lhave "trickled uIP" to hig,her-inicome families. The Kanpur project was Experience in these projects showr's * Take tirh ltel, ro,/ulate nJ, alidt * largely ullable to upgrade slums tle p'it,fills of trlilln to dieal sinililtI7- islstifntitOlll frI7MwrL-clvr'- aiSrl ipitt situatedi on private lands, and slumrr ncaus/v u4ith a broadct ran',4(e of subsectors, 1lt'tlinit ill tWlit, d1qi /r of itaii t?JI 7e t.s. improvement in parts of Madras has problemls, alndt institutionrs ill opte rationis Even the comparativelv sLuccLsstUl led to the eviction of some low- ttiiy eted ioil aindidual citie S or inctropoli- Madras anid KainpuLr projects Fi[-h- income residents. In both cities, hill a7reas. Tlihe nirma/itliv coiplt'x dc'sign light tile] difficultieBs (if tIn/ills to erstfr- project i nfrastructuLre is still poorlv of scluih projects imakes tIlhtem pa1rticimlazrhil emice mIlational or state itiirb inmi p1olic/ maintained, and solid waste services slus;ceptible7 to ImaCrotVIcmlomic instabilitV through interv'et}ions tarL,,v tte(ti ll in some areas are still poorly ruLn. ti/i politieal-institutiomnal chmne. [t also1 siinglc metropolitain area. Phe poota-i ctlls for al7oe-avzentra,n Btilik involvement tial impact of botlh theseprojects vas Policy, cost rec(;overy, anid insti- diuri wn' proecvt iipeincintation as zulcl ais constrained by natio na] aI ntL 5taot tutiona3 goals were Only partially preparation, legislation regulating urb-an lan oi nmet. The Madras project clearly d evelo pment, public Lan,i Li cqutisi- beniefittedi from being a followv-on In view of thie experie nce with tion, and shelter. operation. But it couldi not signifi- integratedi urbani development, more canitly alter state resouLrce allocation recent Bank urbani projects have * Keep project f as itl re Cti patterns for housillg. Nor didI it tendLed1 to focus more on ind1ividiual possilblk; ilidl/idng fil Itiple ann/ its greatly strengthen the metropolitan subsectors suclh as housing or trans- aildt ilnstititntis lisperses c'fifrsts ,,d t planning agency or thie Madras port. Otlhers have provided lines of incrcases risc s mUnLIicipal government. Similarlv, credit for munlicipal development, the Kanpur prOject li,d not mnLchI covering a range of possible invest- * Cainsidenr the act'aiitti' ol-tiit-i- improve tlhe operation of tile local ments but not as part of a predefined location or 1)rogrtmn-tl/ 1e riolnt iic cieri- water utility. package for aniy particular town or tiols oer ic-simngle-locatiom in tc4rInrtcJJ city. Greater attenitioni hias been given projects. The former arelsss vulnor- Banmik porforlnvicc to the administrative cind financial able to changes in localI pcli tica I strengthening of urbani-particularly conditionls. TheyredluLce therisk- ot Oin bhlaiace, Bank perfor-manlce municipal-i usti tutionis, inicreaisinigly poor perftrmna nce [iv dectntrAl iJi ng was bletter in Mlidras and Katnpur as ail eligibility criterion for receiving implementation and sprŽ.idim,, than in Guayacpuil anid hecife, even additiollo1 investment fundis. 1'e project resoLurces over a lOrger thloLu,gh1 the R\ecife project-thle larg- initial experience suggests that these n umt1ber of cities. Tlievy ,in stiinulate est an1d most coinpleX-Colu mUled alpproaches are likelv to vield better collpetition alltmong diffcr-en tt lo