Repot No O9 Road Deterioration in Developing Countries October 1S, 1987 Infrastructure & Urban Development Department Policy, Planning and Research Staff FOR OFFICIAL USE ONLY Docu.ent of th %%rI lank This document has a rEstricted distribution and may be used by repienits only in the performance of their official duties. Its contents may not othrwise be disdosed without W,rod Bank authorization. FOR OFFIC WI ONLY Ron U IIUW Ru US2API COhIhtRES TAMZ OF COU!h!t Glossary of Key It55............................;.. . . ..... 11 , Executive 8uar. ......... 4 .**. .. . iil Cbaptar 1l Th Stan of the Vord .............* 040** t *0*0.,.0 0 1I Chapter 2s Techlical Options and economic Consequences ..000*0000.0e00,.0*.o....**.e............ 11 Chapter 3: The lndtitutio,lal Challenge . .. . . . . 21 Cbapter 4s Pinnlial lequirements . .. . 30 Chapter 5: Policy ConclusLons and Recommndations 42 Statltical Tables: A: Basic Characteristics of Regional Road Networks.......... 50 It Road Networks by Geographic Sego .. 5l C: load User Tames: Their Structure and Contribution to Goverment Revenues in Selectad Countries 54 AnDeX: Explo.ttg Cost-SU fefaive Options for Road nlvestet and VAIntenance 5*..........................- 5 Road Deterioration ln Developing Countries was prepared under the dVrectlou of Louis T. Pouliques, Director, Infrastructure and Urban Developmnt Depatusut. The main contributors to the report were Clell Barral, Asif Isis, Grahm Smith, and Anil Dhadari. sra lBnnathan prolded useful counsel throughout and ssisted in drafting varlow versions of the report. fer Possberg and Vllia. Paterson provlded help on technical and egineering lssues and Udard Holland In drafting the report. The guldace and coments given by S. Shauid Bsin, Pfrwes Hasan, Curt Carneuark and St Alao Walters are grate- fully acknowledgedo. Thnks go to regional tranport staff who reviewed the report at various stages, to meta de Coquereausont and euce Ross-Larson for edting, to Rodrigo Archondo-Callao and Olivier Dottris for reearch asistance, and to Hart Uhokal and ?arjevna Qstrick for typing. ThM document has a stdctd dtbuon and may be us dby eiApiensM on In te puwoasm of th ofcwl dutlL its content ma notmot heww be duciosed wktou Word Seek A d=* - ii - Glossary of Key Terms Road Conditions GOOD Paved roads substantially free of detfects and .requiring only routine maintenance. Unpaved roads needing only routine grading and spot repairs. FAIR Paved roado having significant defects and requir!ug resuirfac- ing or strengthening. Unpaved roads need ng reshaping or z surfacing (regraveling) and spot repair of drainage. POOR Paved roads with extensive defects and requiring immediate rehablbitation or reconstruction. Unpaved roads needing reconstruction and major drainage works. toad Maintenance and Improvement Works ROUTINE MAINTENANCE Local repair of roadway and pavement; grading of unpaved surfaces and shoulders; regular maintenance of road drainage, side slopes, verges, traffic control devices, and furniture; roadslde cleaning, dust and vegetation control, snow or sand removal, and maintaining rest areas and safety appurtenances. RESURFACING Resealing a paved road or regraveling an unpaved road to preserve its structural integrity and ride quality. RERABILITATION Selective repair, strengthening and sh4pe correction of Ipaveent or roadway (including minor drainage improvements) to restore structural strength and ride quality. RECONSTRUCTION Renewing the road structure, generally using existing earthworks and road alignments, to remedy the consequences of prolonged neglect or where rehabilitation is no longer possible. RESTORATION Major rehabilitation and reconstruction works considered together. BETTERMENT Road improvements related to the viAth, alignment, curva- ture, or gradient of road (including associated resurfacing and rehabili- tation works) to improve traffic speed, safety, or capacity. NEW CONSTUCTION Constructing a paved, gravel, or earth r ad on a new alignment; upgrading a gravel or earth road to paved standaro; providing additional lane capacity; or c istructing additional carriageway., front- age roads, grade-separated interchanges, or multilane divided highways. )x cutiv Sinr 1. The developlng 4usttries hae lost precious iftastructure worth billions of dollars through the deterioration of their roads. *f they do not tmudiately begin to do much more to preosorv their roads, they will lose billions more. Large road networks, built at groat expense,, have been under- maitained and more heavily used ad abused tha expected. Uf this neglect continues, the deterloratioik of roads will accelerate as old p_vements cruble and new ones outlive a fairly long period when the effects of "eglect are barely perceptible. .i. hAs cost of restoring these deteriorated roads Is going to be thre to flio time greater than the bill would hav been for timely maintenance. And restoration is only part of the coot. Vehicle operating costs rapidly outpace the costs of road repair as the condition of roaus pases ftm good to fair to poor. Together, tbese avoidable costs are going to be a heavy drag on further economic developmnt. Th tlh4 KSll fotr the eNt 10 teons S 9 Jlln or owasb Iare til. in the 85 counttries tha have received Worlt lank loans for roads, a quarter of the pawed roads outside urban are" ased rdonstruction as do a third of the unpaved roads. This work will cost $40 to $45 billion. (Earlier preventive measures costing les thn $12 billion could have saved these roads and held down the operatiag costs for road users) iv. In addtion, aother 40 percet of the paved roads In tbhse coun- tries require strenogtheAng today or In the next few years. This work, along with routine malntenance, will cost another $40 to $45 billion over the net toeanyers. That brings the total bill for these countries to about $90 billion. But If tbhse countries do not Improve their managemeat of roads, the eventual cost of rs toration could easily lncreas two-to-three fold. And the cost to users would ri-st eve moree Mat cased th leterieratin? v. In the 1960. and 1970s, road networks expanded much faster than the correspondlng maintenance budgets and lnstitutional capacities. Traffic has also becom mch beavler than expected, and axle loadings bave often exceeded the designed capacity of pavemets. These patters -are evident almost every- here. Les evident is what has been bappening to the roads. vI. Ubw paved roads, If Inadequately maintaiaed, deteriorate slowly and almost imperceptibly during their first 10 to 15 years, dependlig on the traffic. After that grace period, the pavemsts deteriorate much more rapidly. Without timely malatenance, they break apart. - iV - vii. As roads become rougher, the costs of operatlng vehicles -- and of transporting goods - begin to shoot up. The neglect of maintenance contin- uses, however, because it is the vehicle operators that pay these costs - and pass them on. Road authorities are sheltered from these coets. Nor do road authorities come under immediate pressure to laprove road conditions. Road users are often slow to see the link between road conditions and the prices of goods and transport services -- and even slower to try to do something about it. Vili. In the absence of public pressure and a clearer understanding of the seriousness of the problem, few governments have given road maintenance a high priority in their budgetary allocationa. The urgency of the situation has not always been fully appreciated by all donors and lending agencies, some having been readier to provide funds for new construction than for maintenance. New construction has sharp political visibility, maintenance little g-amour. Inadequate maintenance in developing countries has various causes, but institutional failure is the only explanation for its wide extent. At the heart of this failure Is the absence of public accountabilityq. ,,be Requirements for 1fflcient Restoration ix. The task now facing many developing countries - as well as the Bank and other aid agencies - is to salvage roads that have deteriorated severely and to protect newer roads from a similar fate. This task has major financial, technical, and institutional requirements. Of these, the institu- tional requirements are the most pressing. X. The techuical options today seem richer. Thanks to advances in understanding the physical process of road deterioration, there now are ways to keep roads serviceable at lower cost than before. 'ihe financial require- nents of road deterioration, on the other hand, are large. Some countries will require large Infusions of external capital; all will need political resolve to mobilize resources at home. But additional financial resources will not by themselves assure effective restoration and preservation of roads. Nor will they prevent a recurrence of the kind of crisis and waste so evident today. Needed above all is a reform of the Institutional base of the road sector. Accountability must be stronger In all entities and activiti3s that maintain public assets. Public and political attention must be delibe- rately fostered. The organization, manning, and activities of the institu- tions looking after roads must be improved to Increase their absorptive capacity. xi. Technical reguirements* Stanadrd engineering practices have very different effecta in different environmeats. Road Investment and maintonance strategies mast therefore be tailored to the circumsta&ces cf Individual countries. The growing understanding of road deterioratlon nevertheless offers some general guidance. xll. For one titing, the savings in vehicle operating costs from paving lightly traveled roads are less than previously eatimted. And the total life-cycle transport costs on paved and unpaved roads are nearly the same over a wide range of traffic volumes (typically 150 - 400 vehicles a day) - so long as the roads are reasonably malitainod. Big differences in cost arise only if aintenance is not done or deferred. If the availability of funds for future upkeep of pavd roads is uncertain, it Is economically safer to keep lightly traveled roads unpaved and reasonably maintained. xiii. It also makes sense - In tmses of budgetary stringency when road allocations are not too far below the optimau - to consider alternative maintenance strategies., Some strategies Involving small cuts In spending barely affect the costs to users, Others, while soiilar in amount, can drive user costs up substantially. Still other strategies, Involving much deeper cuts, can multiply the future cost to both the road agency and the user - so much that they are self-destructive Por example, if the budget cuts are so large that they preclude the resurfacing or strengthening of paved roads now in fair condition, these roads will soon fail structurally &id require much more costly restoration. This, regrettably, ti the situation in maay developing countries, and It will soon sprad to many maore. xiv. HRes are some of the technical cboices that each country should examine in light of its needs and capabilities: G Gravel roads should be pavd only after a complete analysis of the costs, climate, present and future traffic, and reliability of future maintoeace. * If the volume of traffic indicates paving, the strength of the pavements should be guided by the likelihood of undermaintenance and excessive axle loads. It that likelihood is high, pavements should be built to the required strength Immediately rather than in stages. Deterioration of paved roads is gradual and barely perceptible during a long initlal phase that lasts up to two-thirds of the pavement life cycle. Resurfacing and sttengthening can thus be deferred somewhat so long as the roads reain in fair condi- tion and have not entered the final critical phase of their life cycle. Longer deferral will cause pavements to break up. The stage of a pavwd road In its life cycle must therefore determine the decision to defer resurfcing and strengthening operatios. 0 Traffic Is critical for maintenan decisions. If funds are tight, it may be beat to keep heavily traveled roads in fair or good condition and to reduc substantially the maintenance of roads wLth light traffic. Uven If funds are not tight, lt may make llttle economic sene to fully restnre paved roads that are lightly traveled. And if funds are very tlght, it may be beet to let lightly travled roads deteriorate further, applyLig only low-cost patchiLg to keep them usable. - vi - sv. Institutional requIrements. The lnadequacsis of road maintenance stes in part from the structure and functions of the tradltlonal road agency. Often a public monopoly, the agency has too many responsibilities - for plannlng, controlling, and executing both construction and maintenace., And the agency typically devotes too many staff, funds, and facilities to execution - to the detriment of plav2ing, control, and evaluation. It listens only to itself. That makes It desirable In many countries to separate the execution from the other functions, lodging It In a separate government agency or in the private sectoro Where this separation has succeeded, the lncentives for good performance have beon strong, and the delineation of accountability has been clear. Xvti A road agency also needs an effective maagemnt Information system to anictor traffic, road conditions, and (just as Important) Its internal workings.e Th agency should be equipped to analyse the life-cycle costs of construction and malntenance and present and future cost to users. Testing alternative design and maintesnance policies for their sensitivity to differ- ent discount rates, traffic loads, and other variables will be difficult without a specially designed computer model. Without a reliable data base and capable staff, It will be Impossible. xvii* The road agency should, above all, Introduce mechanism to Increase the accountablilty for perfomansce and for resource use - and to provide Incentives for good performance In the agency and by the contractors it egages. It should, moreover, work with the medla and with nongovernmental orgaisaltions to make the political leadership and the public aware of the effects and hlgh costs of failing to maintain the roads. xviii. Financial requirements. The present bill for repair and mainte- nance is about $90 blibion, or f9 billion a year for the next 10 years. That could arrest future deterioration and clear up the backlog of economically warranted restoration. If the work were finished In five years, however, the costs would come down to less than $70 billion - or about $13.5 billion a year. (toad spending In 1984 was about $13 billion, but only a half went for restoration and maintenance.) xixo These aggregate sums obscure wide dtifference from country to country. The 61 developing countries having data on road spending fall Into three categories of capability - with about a third In each. First are the countries that could meet future maintenance requirements and elidinate their backlog In five years with no Increase In their total spending oan roads - but only by holding back on new construction and allocating up to 80 percent of that total to restoration and maintenance. Next are the count4es that could do the saw In 10 years - but only by raising their spending for those years by 50 percent and by allocating 80 percent of that total to restoration and saintenance. Last are the countries that would have at least to double their spending - even If they devote all of It to restoration and aiatenance. XX. To justify more spending on roads and Its reallocation to restora- tionand gmaitenance, each country will have to perform a systesmatic analysis to determine how much should be spent on roads - and how. Some countries may chtwse to revise user charges or impose new ones, which if translated into better roads can reduce rather than increase vehicle operatilug costs. Some may decide that to avoid misallocations they will have to earmark budgeted funds for restoration and maintenance, especially the funds that might be raised through a temporary surchsrge. And some will have to rely heavily on externtal sources of finance. lmlicatious for the Baok xxi. Economic losses - actual or potential - resulting from Inadequate road maintenance can have important implications for the Bank's macroeconomic dialogue with its borrowers, particularly in the context of public expendi- ture priorities. The evidence in this paper suggests that in a number oi countries the economic returns from expanded expenditures on road maintenance would justify a substantial expansion of such outlays relativea .o expenditure on new road construction. It is also quite possible than an expansion in total expenditures on roads rslative to other sectors would be warranted in some countries.. That determination, however, can be made only by a compre- hensive review of public expenditures which evaluates the comparative bene- fits of proposed outlays In all major sectors. xxIi. The developing countries dlffer according to their need to expand and Improve their road maintenance organizations and procedures, their need to Increase their total road budgets and sake reallocations from construction to maintenance, and their need for external assistance. In addition, some have obsolete networks, others new ones. These differences suggest several distinct categories of need for external assistance In the roads sector. The Dank's lending for roads should be tailored to the circumstances In each category - and be conditional on an acceptable distribution of road expendi- tures for maintenance, restoration, and now constructione o Por countries with the best performance (Category I In the text) - largely, but not exclusively upper middle-income countries - the Dank will view itself mainly as a vehicle for institutional and technological lmprovements. It will nevertheless extend funds for balanced programs of maintenance and new construction. * For countries where road maintenance is inadequately funded (Category II) - including a substantial number of middle-income countries - the Dank will provide external finance mainly for maintenance and restoration. Such finance will be linked to policy and institut:ional reforms. F For countries vith severe financial and institutional con- straints on aintenance (Category III) - almost exclusively low-income countries - the Bank's assistance over the next few years will be devoted with few exw,ptions, to maintenance and restoratione The assistance will be conditional on iastitu- tional reforms and on the allocation of road budgets exclusively - viii - to maintenance and restoration (with only llmit.d and defined exceptions). For countries with obsolete roads, technology, and organization (Category IV), the Bank will emphasize organizational and tech- nological improvements In modernizing the road networks and will also support new construction when it is justified. In all this, the Bank will encourage the execution of maintenance work by entities outside the road authority - entities operating on commerclal prin- ciples, preferably in a competitive environment. In addition, the Bank will promote measures to educate officials and Increase public awareness of the need for timely maintenance and the high cost of neglecting roads. xxil. The Bank will also promote more coordinated action by the intern.- tional community ln dealing with the growing problem of road deterioration in the developing countries. To this end, the Bank will work closely with other aid agencies in designing road programs appropriate to the necds of each country (along the lines of rhe four categories just spelled out). The Bank will also work closely with other aid agencies In supporting research on - roads - work that would be done mainly by road agencies and road research institutions in the developing countries. The emphlsis In these efforts will be on international exchanges of data, technology, and management information systems. The Bank will support compilation and publication of statistic6 on roads - work that should be undertaken by the United Nations Statistical Office or an organization such as the Permanent international Association of Road Congresses or the International Road Federation. External agencies should provide financial assistance for this effort. xxiv. The recom_endations for World Bknk policy emerge from a view of the root causes of road deterioration in countries where It Is now severe. We know now that some past investments have been mistakene We know, too, that deterioration reveals Its symptoms late and gives little warning of what Is to come. Added to these factors are the civic calamity in some countries and the harsh and prolonged economic adversity In any more. But where the road problem is now serious or is about to turn serious, the damage attributable to such other factors could usually have bemn contained by more effective, responsible, and attentive Institutions - by more efficient institutlonal arrangements for the management of the cvtintxy's infrastructure. It is diff- icult to understand fully the causes of institutional failure in different countries - even more difflcult to prescribe countermeasures guaranteed to cure the problem. But this is no case for leaving things as they are. Action is required, even If it has to be accepted as experimentation. Bank support for the road sector of countries with serious road deterioration problems will therefore be guided by signs of Institutional progress These signs Include the growth of political attention to the preservation of Infra- structure, the strengthening of accountability in the institutions charged with managing it, the introduction of Incentives that press people to per- form, and the deliberate search for resources and schemes that promise care for hard-von national assets. CUPU 1: oTE T oT lOA 1.01 The developing world's road building boom in the 1960. and 1970. created an infrastructure that has been crumbling In the 1980s. If not quickly strengthened and protected, this crumbling infrastructure threatens to collapse in the 1990s. Large road networks, built at great epense, haze been inadequately maintained and used more heavily than expected. The result in many developing countries ls a 4etwork of deteriorating roads. Many roads are in such poor condition that normal maintenance to no longer sufficient or effective. These roads now require rehabilitation or recon- struction at three to five times the cost of timely preventive maintenance and strengthening (see the glossary of terms). And many more roads, whose deterioration is not yet visible, will soon reach that point without sub- stantial improvements In maintenance and rehabilitation. .02 The problem of poor maintenanes are worse for roads than for other sectors for three reasons: The costs and financial requirements are very lare. For example, expenditures to compensate for past omissions in preventive main- tenance in Sub-Saharan Africa are at least 10 times those of supplying textbooks for all elementary school children in the region until the year 2000. R Boad deterioration accelerates with time. This phenomenon makes It difficult, but also al the more important, to recognize the need for preventive waintenanee before deterioration becomes obvious and more expensive rebabilitation or reconstruction becomes necessary. *Road authorities are Insulated from the effects of undermaaiute- nance. The agency responsible for road maintenance usually Li not exposed to either the economic consequences of underaintenmnce or This report assesses road deterioration in the 85 developing countries receiving highway sector assistance from the World Bank Group. [111 The analysis covers only the main road networks, both because they are the most important and valuable roads and because informatlon on other roads is too fragmentary. In these countries, 70 to 80 percent of Interurban traffic is on the main networks. These networks consist of 1.8 million km of primary and secondary roads, of which 1.0 million km are paved, and have a replace- ment value of about $300 billion, excluding the value of land, bridges, and major earthworks (See Statistical Table A). The main networks generally include the principal roads and highways crossing urban areas or providing acceos to ports. The analysis does not deal with urban roads; nor does it cover 5 to 6 million kllometers of local rural roads and tracks (mostly upaved and with low traffic volumes), with a replacement value of perhaps $75-$100 billion. -2- the orgasixed pressure for better roads, and the urge to provide responasive, effective maintenance ts thus weakened. 1.03 The costs to road authorities are only the tip of the Iceberg, for the costs to road users operating vehicles on rough roads are much larger. These high and rising road haulage costs constrain the location of economic activity, hmper the Integration of economic markets, limit the gais1 from specialization, and reader unviable many activities that rely on transport (See Box 1.1). Am Overview of load Coudititos 1.04 Detailed assessments of road coanditions at present are subject to considerable error. The Importance of systematic monitoring and evaluation of road conditions has only recently become generally recognised. Further- more few developing countries have technically adequate data bases and management Information systems. Nsverthele,ss the evidence io good enouSh to show the broad dimnsions of the problem. Receut field surveys supple- mented by the judgment of World lank engineers, permit the broad categoriza- tion of the distribution of a country's roads mong three elases of condi- tion: good, fair, and poor. C23, 241 A road In "good" condition requires only routine maintenAnce to reain that way. A road In "fair" condition needs resurfacing. A road In "poor" condition has deteriorated to the point that It requires elther partial or full reconstruction (See Glossary Lor a definition of road maintenance and improvement works). 1.05 Three facts about road deterioration help to clarify the problea. First, becau"e reconstructlon costs three to give times "smuch as resurfac- ing or rehabilitation, no road should be allowed to decline Into poor condi- tion urless it Is to be kept deliberately In that condlton (with routine maintenance but no resurfacing or rehabilitation) or abandoned entirely. Second, paved roads In fair condition normally have a window of opportunity of about five to eight years when they can be restored by reurfacing or strengthening. The existence of many roads In fair condition suggnets, therefore, that extensive maintenance is needed quickly If these toads are to be saved before they reach poor condition. And third, the cost of operatlng vehicles (especially larger trucks) rises as roads deteriorate. Because vehicle operating costs are the largest part of road transport costs on all but very lightly traveled roads, the increse In operatlg costs swamps all other costs at stake it road maagement as oads deteriorate. 1.06 The road conditions In different countries combined Into regonal averages, are alarming on two counts (Table 1.1). First, more than a fourth of all paved roads - some 269,000 km -- are already In poor condition and need rehabilitation or reconstruction.l/ See-ond, the heavy concentration I/ Less serious dAeterioration In the United States (US) federal ald highway network during the 1970. prompted widespread alarm, new legislation, new user taxes, and a large infusion of federal and state resources in the 1980. to remedy the situationa [2, 401 3 Bo0 1.1 o O f bad Nsel in 6w.. Rad deterioration can make an economic crisis worse. In Ghana a good road network ws built before 1970, but it later suffered from serious anglect. By 1984, about 60 percent of the main paved roads were in a state of moderate to severe deterioration. Important sections have become almot lmpassble, and access to some Interior areas has been severely curtailed. Transporters refuse to go there because they do not want their vehicles to break down. Transport costs have increased In real terms by about 50 percent oan mainroads and by over 100 percent on rural roads that have suffered even greater neglect. In many areas, the market rate for trans- porting fertiliser iS as hiSh as a dollar per ton-rile. These high transport costs have cut deeply iato farm returns, particularly for poor farmoe In areas away from main roads: same villagers can no longer move thelr cocoa stocks to regional depots. And during the 1982-83 femine, poor roads prevented the transport of food from surplus areas to areas facing starvation. The high transport costs have also hit the timber Industry. Logs moving from the Kumas area for export through the Takoradi Port are trAcked over a 500-kilometer route because the direct road, whicb is about half as long, is broken down and the rail services are unreliable. The detour adds U8$15-20 for each ton of logs exported. The prospects are not encouraging. Road resurfaclng and strongthenAng covers only about 20 percent of estieated anoual needs of about $47 million, mainly for z"esling 1,100 km of paved roads and regraveling 1,600 km of unpaved roadsd *ven wlth a recent road mainte- nance project supported by the World Bank, maintenance will lncrease only to about 30 percent of the noed by 1989. Unless more resources can be put into road maintenance and applied effectively - deteriation will coatinue. Ultiately It will require far more costly rebabilitation, If financed, or result ia an almost total collapse of road transport. -4- of paved roaids in fair condition (42 percent) foreshadows a major crisis unless concerted efforts prevent these roads from deteriorating into poor condition. 1.07 The unpaved roads in the main networks are even worse. True, a road authority's timing of maintenance for inpaved roads is less critical than for paved roads because the costs of restoring unpi.ved roads are less sensitive to the tiaing of intervention. Even so, the aggregate effect of thelr deterioration on vehicle operating costs can be considerable. Although unpaved roads normally carry much less traffic, the riding condition of an unmaintained unpaved road deteriorates any ti&As fuster than that of a paved road (except in the terminal phase of the paved road's life cycle). Defer- ring simple routine maintenance on unpaved roads can quickly double the vehicle operating costs. Najor Determinants of loAd CndNitions 1.08 The considerable variation in road conditions In different countries and regions stem from differences in the past maintenance needs of individual networks and in the countries' responses to those needs. 1.09 Need for mintenance. The maintenance needs of a road network can be predicted fairly accuratelry from a set of structural characteristics, such as age, climate, traffic, design standards, construction quality, and subse- quent maintenance. Of these, age, traffic, and construction quality are particularly important for understanding the situation in developing countries. 1.10 Age is important to the condition of paved roads because of the time-path of their doterioration. Typically, two-thirds of pavement deterio- ration (and an even higher proportion of maintenance cost) is concentrated in the final third of the design life of the pavement (para. 2.03). After a boom In road construction, a grace period of several years - when roads remain in good condition regardless of maintenance - is followed by a period when the need for maintenance surges dramatically. 1.11 Differences in the age of networks underlie the regional differ- ences shown in Table l.l. In West Africa, paved roads are on the whole in better andition than elsewhere. The networks there are fairly new: more than half the paved roads were constructed, upgraded, or reconstructed during the last 10 years* Nigeria and Cote d' voire have substantially rebuilt their paved networks since 1975 (See Box 1.2). Some other countries, less well endowed, were helped by external aid agencies to develop and improve their road networks. Of 20 West African countries for which information is available, 15 have young networks. The proportion is not much lower in Eastern or Southern Africa, where 10 of 16 paved networks are fairly new. Without substantial external assistance, many of these networks are unlikely to remain serviceable beyond the next 10 years. 1.12 The growth of traffic on roads built to obsolete design and construction standards helps to explain the general road conditions in South . Table 1.1 Condition of Main Roads a/ (percentages weighted by7length of country networks) _aved Unpaved Region Good Fair Poor Good Fair Poor Eastern Africa 42 32 26 42 30 28 Western Africa 52 23 25 20 36 44 East Asia 20 59 21 41 34 25 South Asia 19 45 36 6 39 55 Europe, Middle East, and North Africa 41 35 24 30 46 24 Latin America and Caribbean 44 32 24 24 43 33 Average 32 42 26 31 36 33 United States (Federal Aid Network, 1981) b/ 31 57 12 United Kingdom (Trunk Road System, 1983) / 85 '12 3 a/ As reported in an internal World Bapk survey of 85 countries based, as far as possUble, on published pavement condition information (60 countries) and supplemented, where necessary, by the judgment of Bank highway engineers. See the glossary for definitions. b/ After Public Works Infrastructure: Policr considerations for the 1980'8, pp. 20-21. 1401 c/ Adapted from Imnestment In the Public Sector Built Infrastructure - Report A: Roads and Brie p0. [381 -6- Box 1.2 Niger'a's bw bDaes .d the Risk of Massive Det derti.a Of the 21,000 km of paved roads In the Nigerian federal Trunk Road System, 36 perceat vero constructed or rehabilitated during 1975-80 and another 24 percent during 1981-85. As a result, 62 perqent of the aetwork Is rated good, 15 percent falr, and 23 percent poot. Neglect of the unpaved network has resulted In an est mated 90 percent in poor condition. Despite the remarkable expansion and lmprovewmnt of the Nigerian trunk road system during the last 10 years at an,estimated cost of about US$$ billion, the situation is still precariti4. Many of the roads were built with generous geometric features but imak pavemets, which require substantial strengthening. Wtthout an extesive pavement resurfacing and strengthening program covering about 2,000-3,000 km a year, at an estimted annual expenditure of about $150-200 midliio, the Nigerian trunk road system may deteriorate rapidly, requiring massve rebabilltation and reconstruction In the next 10 to 15 years. -7- Asia: only 19 percent of paved roads remain In good condition, and 36 percent are in poor condition. South Asia's road networks consist mainly of aging roads with neither the geometric capacity nor the structural strength to carry current traffic. Both the volume of road traffic and the axle loads have increased there over the past decade as economies have grown and traffic has shifted from other nodes. Por these networks the replacement and upgrad- Ing of obsolete roads Is needed beyond normal maintenance and rehabilita- tion. Countries in South Asia, particular.Ly India, face the need for a very large program of road building. 1.13 The age of networks and the growth of traffic explain some diffea- ences In maintenance requirements, but the condition of tb.i roads reflects the extent to which maintenance requireents have been met. Where mainte- nance has failed, It is largely an outcome of misallocated funds, unsound aintenance strategies, and Inefficient Implementation. 1.14 Financial caaclit. The capacity to pay for road maintenance from domestc sources depends on a country's resources. Gross Natlonal Product (GNP) per capita may thus be a good Index of financial capacity, although some governments are able to capture more of It than others. And of the fiscal Intake, the allocation to highmays Is a political decision, reflecting a judgmeat about national priorities. Variations In Income growth also affect funding: severe setbacks In Income growth clearly explain some Impor- tant instances of underfunded maintenance and consequent road deterioration. 1.15 load conditions in Latin America are worsening rapidly, mdrroring the general economic dovnturn and accopanying financial stringencies over the past five years - and Illustrating bow qulckly roads can deteriorate once tbey reach the critlcal age. Brasil, with by far the largest road net- work In Latin America, has seen serious erosion despite substantial efforts. In 1984, 28 percent of the network was in poor condition, up from 18 percent in 1979 (see Box 1.3). The deterioration was even more dramatic in Honduras, which previously had adequate road maintenance. The percentage of the paved roads In good condition there dropped from 82 percent In 1981 to 50 percent in 1984, due largely to flianclal constraints. 1.16 A broad Indicator of the economic burden of road malatenance is the ratio of a country's road network length to gross national product (GNP), with an allowance for the lower traffic volumes In poorer coustries. The ratio for main roads ranges from 0.3 km per million dollars of GNP for Korea and 0.4 for Nigeria to more than 8.0 for Zaire and Botswana and 14.8 for Quinea-Bissau. Of the 36 countries with the highest ratios, 32 are In Sub- Saharan Africa (from Statistical Table B). *ben with the best manageaent, these countrles face the world's highest burden of road maintenance require- ments relative to lncome and are likely to have lower than average mainte- nance. 1.17 Financlng road maintenance by borrowing from external comercial sources abroad or from developmet institutlans could help to fill the funding gap, but many of these sources have preferred to finance construction rather than maintenane. x 1*.3 A Cam of i at Ir ll's federal N1 tuok A 1979 survey of Brazil's federal highway network showed the following distribution of road con.ditions: Ood 10,000 km 242 ailr 23,000 km SaX Poor 7,000 km 18 Restoration and preventlve maintenance of all roads at that time would have cost $1.8 billion Zn constant prices. A repeat survey In 1984 rated the network as follows: Good 14,000 km 302 Fair 19,000 km 422 Poor 13,000 km 282 The Increase In the percentage of roads In good condition resulted from neW construction not good maintenances 6,000 km of new paved roads were constructed, while 2,000 km of those formrly In good condition declined to fair condition. Nor did the masive backlog of roads in falr condition receive the resurfacing and strengthening that was due. Some 6,000 km of that group deteriorated to poor condition, greatly Increasing the number of kilometers and the percentage of roads needing rehabilita- tion or reconstruction rather than just preventive maintenance. The cost of this restoration Is estimated at $1.7 billion, and the cost of preventive malatenance needed to save the roads still In fair condition Is estimated at about $750 million. Thus the federal rehabilitation and maitenane backlog gtre to $2.4 billion, a one-third lnerease In six r~~~yas - 9 - 1*18 Institutional capacity. The forogoing Indicators - the differ- ences in GNP per capita and Its rate of change, the ratio of network length to GNP, the age distribution of roads, and the volume of traffic - do not fully account for the variations In road conditions ftom country to country. Several higher-income deweloping countries have poor roads while some of the lowest income countries have better ones. The capacity of a country to deal with Its maintenance needs also depends on how effectively and efficiently It can translate funds Into protecting and rehabilitating the road Infrastruc- ture - generally known as Institutional capacity. 1.19 Institutional capacity has several 'acets. one Is the pool of skills, such as the slse of the labor force that can be applied to the acti- vity. Others are the soundness of the maintenance strategy - type, level, and timing of Intervention - and the managerial and operational efficiency In executing the strategy. 'These depead, In turn, on such factors as govern- ment commitment, Institutional structure, managerial abIlity, staff quality, accountability, and Incentives. Experience with these aspects of lnstitu- tional capacity has often been remarkably disappointing. And where limited Institutional capacity cannot readily be expanded, the prospects are poor for recovering Infrastructure that has already deteriorated or for preventing. heavy losses of capital where the peak of maintenance needs is yet to coie. The lBrd Core of the Problem 1.20 Not all countries facing a maintenance crisls have arrived there by the same trajectory. Some have backlogs of maintenance needs because their financial and Institutional capacities have not expanded as fast as thelr road aetworks. Some bave built up backlogs by deferring needed maintenance during adverse economic conditione. And some do not appear to have backlogs now, but their mtworks of relatively recent construction will soon require greatly lncreased and systatic maintenance to prevent rapid deterioration. The road deterioration problem thus pervades the developing world. At the core of the problem are mainly the countries of Sub-Seharan Africa and South Asia. And at the bhrd core of the problem are Sub-Saharan African countries whose financil and institutional capacities are unequal to the task at hand or to the one they will son bave to face. 1.21 Altbough some networks in Sub-Saharan Africa are among the better maintained (in Cote d'Ivoire, alawli, iger, and Rwanda), they are also of recent vintage and will soon require mch more mainteunace as the paved roads pass Into the critical stage. In the rest of Sub-Saharan Africa, recently expanded networks are also numerous and therefore In better condition than older ones. In 10 countries with older networkes, an average of 44 percent of their paved road length Is In poor condition: in the 17 countries with fair- ly new paved roads, the proportion of poor roads Is only 22 percent., There l, bowever, no reason to bellove that countries with naewer paved networks have better capacities or policles for road mintenance than those with older networks. If they do, their unpaved roads should be In better conditiou -- without maintenance, unpaved roads deteriorate rapidly and at a uniform rate. The condition of unpaved roads in countries with newer paved roads - 10 - differed little from that in countries with older paved roads. So, the state of a country's paved roads says little about Its road maintenance capacity. 1.22 In many countries, the rapid expansion of networks has outstripped the growth In Institutional capacity for road maintenance. But full conse- quences of this lag are not yet evident In most paved networks, because most of the roads are still in the grace period when they need little maintenance and show few signs of deterioration even when maintenance is neglected. The effects of undermaintenance are already reflected In the relatively poorer condition of unpaved roads In all regions. In Nigeria, 62 percent of the paved roads are In good condition, while none of the unpaved roads are rated as good. In Cote d'Ivolre, 78 percent of the paved roads and only 30 percent of the unpaved roads are classified as good. Both countries developed their paved networks fairly recently, so their maintenance requiremeats will grow significantly In the next few years. Almost every region has some cauntries In or approachlag crisis, but Sub-8aharan Africa stands out as having the most countries at the hard core of the road deterioration problem. cHAPm 2: UE OMOU AM ECOICIC 2.01 A few fundamental relationships explain road deterioration and its consequences for the total costsof road transport. They hold in a wide variety of environments and form the starting point of any rational malte- nance plan. Recent research has greatly clarified these relationships and has quantlfled and refined them empirically, thus enabling the design of economical maintenance strategies for a wide range of circumstances. And they show that some past strategy decisions have been wrong. Above all, these relationships make it possible to provide firmer guidance for countries that have to restore badly deteriorated parts of their networks while keeping the other parts serviceable. 2.02 Economic decisions about a highway system must take into account the totRl cost of transport on the roads: the discounted life-cycle cost of constructlg and maintaining the roads and the (usually) far larger cost of operating vehicles on these roads. The estimation of these costs, in turn, mut be based on sound knowledge about road conditions, about how those conditlons and the various vehicle types and operations affect each other, and about the applicability of different maintenance techniques in different environments. noe umaatalUmatioms 2.03 As roads deteriorate they get rougher. Until recently roughness was assessed subjectively; now It can be quantitatively evaluated using measures such as the International Roughness Index (IRI). l33, 341 Unpaved roads, If not maintained, deIeriorate very rapidly, at a fair'y uniform rate throughout their life cycle. / But uAsaintained paved roads follow a distinctly nonlinear path (Ftgure 2.1). During a long initial phase that lasts up to two-thirds of their life cycle, paved roads undergo little visible deterioration. This stage is followed by a phase of Increasing deterloration - slow at first but then rapidly accelerating (into fair condition) - that ends within a few years In radical structural failure (poor condition). This noullnear deterioration affects the choice of optimal maintenance policy for paved roads. For the unwary, it also tends to disguise the future maintenance requirements of young networks. 2.04 During the first phase, maintenance needs are minimal, and a paved. road can be kept In good condition by fairly inexpensive routine maintenance. 1/ All references to the absence of maintenance assume that the minimal work of vegetation control and clearing draiasge is still carried on. These activities, which are very low in cost, are vital to the coutinulng serviceability of roads, and neglecting them Is. a sure way to hasten road deterioration. - 12 - npi 2.1 Road Deterioration Over Tim S~~~~~~~~~~~~~~~~~~~~Se * - (Il##9w . XS iiia7UJ t f w a o P + |- -t w w s n- wa t pw 7 vw vwww-w ftltl nz t o F 0 lb ngm 2.2 Votdcle Operatlwg Costs as P samtap of Total lfs.Ccle Sramport 00tSU for 7yIcal Cases wdtb Opt11 mtlFlate ~~~~~~~~~= = M dtial Avrwe Ibtly Traffic (32 &==I traffic Scoth rate) btens PrEOa valus di sconed at2 percent a ye£or construction + maintenanc vehicle operating cost e - 13 - In the subsequent phase of Increasing deterioration (into fair conditiot), the pavement 4' be restored to good condition. This can be done by resur- facing or by strengthening the pavement with an overlay - a moderately costly measure that restores structural strength and ride quality to met traffic requirements for the next siveral years, thus starting a new paveent cycle. With adequate routine maintenance and timely resurfacilg or strengtb- ening, a paved road should never deteriorate Into poor condition, which requlres far more costly reconstruction. 2.05 A second fundaaental relation links vehicle operating costs to road roughness (Table 2.1). Recent research has shown that the effect of rough- ness on vehicle operating costs is not as large as previously estimated. In so far as comparisons are possible, the new estimates of the Increase In operating cost as a function of roughness are on average a quarter below those proposed In an earller standard work (see Annex, pars. 3-4). Over- estimates of the savings in vehicle operating costs may have led to mistakes In Investment and maintenance decisousi, but a far graver error occurs when maintenance policy Ignores the share of the vehicle operating costs In total road transport costs. That share is high, except where traffic is extreely light (Figure 2.2). Even a small percentage Increase in vehicle operating costs from rougher roads Is significant, outweighing the cost savings from deferred maintenance. Strategies for toad Imvestmmt and Naintemsae 2.06 Much Information about the basic relatlons governing the process and effects of road deterioration comes from empirical research supported by the World Bank (Box 2.1). The large data set and the relations obtalned from It are Important contributions to road strategies for developing countries and to policies that eancompass both construction and maintenance decisions. A model based on that research has been used In studies of road maintenance in several countries. For Mali, Chile, and Costa Rlca, these studies demon- strate the interdependence of Investment decisions and maintenance actions and the effect of the quality of lnformtion on predictions* (The model and the studies are described In the Amnex.) These countries were selected to represent a wide range of transport conditions In the developing world. The results, though specific to the conditions In these countries, Illustrate general points about selection of cost-effective maintenance policies and the decision criteria for paving and initial paveent strength. An annual discount rate of 12 percent Is used for the analyses reported In thls chapter. (41 2.07 Traffic and road conditions. In establishing standards for road design and maintenance, the key factor is the level of actual and expected traffic. For Chile and Costa Rica, which have relatively higher trafflc volumes on their road networks the optimal policy was to keep the most heavily used two-thirds of the networks In good condition while maintainlng the other one-third at lower standards. In contrast, for Mali, where traffic volumes are generally very low, the most economical solution was to keep - 14 - Table 2.1 Effect of koad Roughnes on Vehicle Operating Costa (Economic Costa, Costa Rica, 1984) Index of Vbeicle Operating Costs (GOOD a 100, at ISU of 2.3) Road Condition (Roughness, IRU)I/ Vehiele class FAIR (4.6) POOR (6.9-9.2) small car 106 114-126 8us 104 109-116 Light diesel truck 111 124-138 Heavy truck 114 129-146 Articulated truck 112 127-144 I he dominant effect of road condition on vehicle operatlg costs is tbrough "rougbhess," a msaure of road surface irragularities stand.ardised by the International Roughness Index (IRL). Source: From application of the veicle operating costs submodel of the Highway Design and Matieance Model. [411 For the Costa Rican context, see 141. - 15 - Bos 2.1 Mhe ligbhuy Design aDd Mintenance Standads Study In 1971 the World Bank lnitiated what later became a major collabo- rative program of primary data collection and research -- flrst In Kenya .and later, on a much enlarged scale, In Brasil and India. With partici- pation and funding by several other institutions and goveraments, over $20 million was spent for data collection and analysis to provide rigorous quantification of key relationships. More than 90 percent of the funds came from the other participants: the governrmets of Brazil, India, Kenya, and the United Kingdom, as well as the United Natlons Development Program. The governeats of Sweden and Australia provided techrAcal assistance. Results of the studies are being Issued as a series of World Bank publications, including two volumes on road user costs (6, 411, a volume onk road deterioration 1271, another on the resulting planniag model 1421, and two technical papers on measuring road roughness 133, 341. This research advanced knowledge about basic road deterioration and vehicle operating cost relationships beyond the state of 10 or 15 years ago. It also revealed the complexity of the relationships and their sensitivity to an array of area-specific circumstances. The relationships were formulated on the basis of a larger and more comprehensive data set than was previously available. To apply tbhs empirical knowledge to the essentially Interrelated questions of highway construction and maintenance, the Bank developed the Eighvay Design and Maintenance Model (UDK) and a companion Expendi- ture ludgeting Model (EN). These models are designed to search for the best solution appropriate to a country's circuestances. They do this by exploring multple combinations of alternative road designs and maintenance options In terms of their effect on total transport costs. The results are then tested for their sensitivity to variations In Input parameters and future conditions - a necessary precaution. The appli- cation of the model nevertheless requires a rellable lnformation system and capable staff who can make proper use of the data base and the mdl. a ~ ~ . . . - 16 - about one percent of the network in good condition, maintalning the rest at considerably reduced standards as determined by traffic. 2.08 Under these optimal policies, keeping roads in less than good condition does not imply neglecting maLntenance. In all cases, drainage and vegetation control are fundamental, as are high standards of patching and basic routine maintenance, even for roads to be maintained at low standards. In Mall, this maintenance is estimated to cost about $6.2 million a year - about twice the current spending. (This estimate does not include the $9 million required for the backlog of economically warranted rehabilitation and reconstruction of higher-volume paved roads in poor condition.) 2.09 To pave or not to Pave. Decisions about whether or when to pave an unpaved road are subject to many factors, one of which Is the expected future growth of traffic. Figure 2.3 shows, for a specific example In Mali, the net present value for paving a gravel road as a function of trafflc. The exact traffic level at which it becomes optimal to pave the road is very sensitl7e to the expected rate of growth of traffic as shown by Case 1 (no growth in traffic) and Case 2 (6-percent annual growth). It is assumed In all cases that proper maintenance will be carried out whether the road is paved or not. The cost of paving is also Important: in Case 3 (no growth In traffic but a 50-percent Increase In cost), the optimal traffic threshold for paving jumps from 270 vehicles a day to 425. But the penalty for not paving at the optimal traffic level is not very large. Considerations other than traffic say thus govern, the choice. 2.10 Ihen the decision to pave is not preempted by present and prospec- tive traffic, it has to be made from a consideration of other factors. Important among these factors Is the probability that future maintenance will be done to acceptable standards, on paved and unpaved roads. If the main uncertainty Is about the availability and adequacy of future maintenance funds rather than the ability to plan and execute maintenance, the best policy will be to defer paving. The presgant value of the total life-cycle transport costs associated with a gravel road (even though suboptimally maintained) will be less than that for a paved road If there is uncertainty about the availability of funds at the stage when paved road deterioration begins to accelerate steeply. Conversely, If uncertainty mostly surrounds the country's maintenance capacity, early paving (and therefore, fever roads) will be Indicated. The present value of the total life-cycle transport costs associated with an unmaintained gravel road will exceid that of am umain- tamned paved road down to the point at which tbey become les serviceable. These conclusions, stated In broadest terms, result from the joint effect of discounting and the nonlinear deterioration of paved (but not unpaved) roads. 2.11 Strong or weak a_vements. An important question in determning the paving strategy Is whether the staged construction of pavemeats makes econo- mic sense: that Is, beginning with a low-strength desAgn and strengthening the pavement later to accommodate heavier traffic and axle loadings. The answer depends on the quality of maintenance that can be expected on the (initially) lower strength road. The question was explored for condition In - 17 - Mall and Costa Rlica, assuming 3 to 4 percent annual growth of traffic and two levels of axle loads. With a 75-percent to 90-percent probability of adequate maintenance, staged construction would be economical for initial traffic flows up to 2,000 or 2,500 vehicles a day, epeanding on axle loads. But with only a 20-percent probability of adequate maintenance, strong pavemeats should be built initially, despite the higher construction cost, to compensate for unreliable maintenance - staged constructiont In this case, would remain the preferred alternative only at traffic volumes under nigre 2.3 Breakeven Traffic Volumes for Paving a Gravel Road. ~~~~100~~~~~~ 0 ' 150~~~~~~~~~~~~~~~~~~~~0 _ ." * ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 0~~~~~~~~ 4. Ca. 2,' a~ ~ ~~~~~~~~o I* 50~~~~~~~~~~1 8 i~~~~lta Avrg Dafiy Trfl vbe,ady * 4 3 Si Ca. 15X^ 0' _ -50 0 100 200 300 400 51 Initial Averag Daily Traffic (vehicles a day) - 18 - 1,000 vehicles a day with light axle loads. In all cases pavement deslgns should conform to minimum design and constructon specifications.!/ 2.12 Some countries made expensive mistakes by staging the construction of pavements under assumptions of adequate maintenance and restricted axle loads, assumptions that proved invalid. Heavy loads accelerate road deteri- oration, and even with optimal maintenance policies, the average life-cycle roughness remains higher with heavier axle loads. Empirical evidence shows that pavement damage lncrease exponentially (to the power of 4) with in- creased axle loads. But the regulation of axle loads has proven exceedingly difficult and expensive. Many road authorities are now building stronger and more expensive pavements than would be necessary with effective axle load control. Even if it were practicable to enforce axle-load regulations, axle load limits should be increased beyond the prevailing 8-10 ton single (and 13-16 ton tandem) axle loads - because the economic gain from the use of larger trucks outweighs the increased damage to roads. 1321 Tactics ugder aodgetaz CoDstwsimts 2.13 Road authority funding often is not enough for carrying out the. economically optimal level of maintenance. And when budgets are constrain- ed, the best policy is not simply to reduce all categories of maintenance spending equally, as is often done. The situation calls for revising poli- cies and using different maintenance options. The management of retrench- ment (and, in extreme cases, network contraction) has not received much analytical attention. But some rules can be stated to control the retrench- ment process and prevent it from becoming haphazard. 2.14 In establishing spending priorities, the tradeoffs between road agency costs and net benefits of different maintenance strategies are not normally examined. As long as the net present value of a maintenance option- is positive, it is deemed justifiable - whatever the cost to the agency. But if the analysis is extended to road agency expenditures, there appears a big margin for reducing them without making similarly large cuts in the benefits to road users. Pigure 2.4 shows net present values for alternative maintenance options applied to a specific class of roads. The line connect- ing the points that are highest in net present value for given agency expenditures is the efficiency frontier. The optimal strategy, given unlimitr. funds at the assumed discount rate (12 percent), would be that shown at point I. In addition to basic routine maintenance, this strategy would require an i _ediate application of an 80-us overlay, followed by 40-mm overlays whenever roughness exceeded 3.5 III. 2.15 Note that the efficiency frontier is rather flat iusediately to the left of the optimum point - and that alternatives in this zone can 2/ A technical option that has not been sufficiently explored in developing countries, particularly in the humid tropics, is the use of portland cement concrete pavements. Where properly designed and constructed, concrete pavements, though Initially expensive, can provide a long initial service period (15-20 years) that is nearly maintemance freeo - 19 - Pigore 2.4 Net Present Value Versus Agency Expenditures for Alternative Maintenance Strategies for a Paved Road in Fair Condition *oo Average daily tr,affic 800 vehicles 42 3I (C) 400- 400 40- 20 so Pawmntt KUtcenanca (A) 0 ~ ~~t 3 050 0 tO 20 30 4 Preent Value of Road Ageny £xpndltures ('000 US$ per hk) Note: See the Annex for specifications of the maintenance alternatives represented by each point on the graph. It is possible to operate on the frontier between two of the points by using one policy on some roads and the other policy on others. For example, a combination of policies 7" and "G" could use funds equal to those required for policy "H3 and achieve a aet present value on the frontier. Thus policy E" is Inferior and should never be used. Similarly, any other points below the efficiency frontier represent policies that should not be used. In the circumstances of the ca& illustrated above, policy D (reseals) is an Inferior solution, compared to a combination of patching only (B) and overlay (7) options. In other cases, resealing will often turn out to be an efficient maintenance measure to prolong the functional life of pavements. The relative position of various policies will change significantly with changes In discount rate and the relative cost of different maitenances alternatives, - 20 - significantly reduce agency costs with only a s ml impact on total traom- port costs. As agency expenditures are cut back, vehicle operatikg costs Increase, but by only a little more than the savings In maintenance epndi- tures. Nbreover, the tradeoffs in the region near the optima do not Imply higher future costs for road rehabilitatione It Is simply a matter of main- taining the roads at a somewhat roughor levels with a consequent saving to the road agency and somewhat higher costs to road users. In general, this sort of tradeoff was found In the vicinity of the optimal strategy for most other cases studied. 2.16 But as the budget departs further from the optima and maintenance Is cut back correspondingly, vehicle operating costs lncrease by much more than the saving to the road agency. In addition, the changes In mainte- nance policy entail n1creasligly greater risks of pavemeat failure, leading to much higher costs for reconstruction in the future. on sum, the range of good malntenance policies Is wide, but below a critlcal level the conse- quences of further reduction becom destructive. 2.17 Sometimes a goverment must take austerlty measures to deal with a national emergency or adjust to difficult economic conditions. Normal bud- gets may have to be cut back for a year or several years, with the Intention of restoring them later. In such clrcumstances, road maintenance Is often deferred. If roads are In good condition, the 4eferral may not be too costly. Increaed rougbaess due to undermaintenance will lncrease vehicle operating costs somewhat, but roads that decline to fair conditioa can be restored with a mAodest additional expenditure. If, bowever, the deferral continues too long or If roads are In only fair condition at the start, the Impact on vehiele operating costs will be greater. And paved roads declin- ing to poor condition will require far more expesive reconstructlon later on. That is why road agencles shoult consider future budget Interruptions when making cboices about tbilr design and maintenance policy. And one way to limit the adverse impa.t of future Interruptions in funding Is tbrough good maintenance In nonrml times. C8LPTU 3: TO I ISI8ZlUTIOIAL (CZLLEEG 3.01 Choosing the best maintenance policy will remain an academic exer- cise until institutions can efficiently put the policy Into praictice. Exper- ience has been diaappointing. The World Bank advanced more than $1.2 billion between 1971 and 1985 for training, technical assistance, and management consultancies to improve the organization and management of road administra- tion in developing countries. Other lenders and donors and the developing countries have also devoted substantial resources for this purpose. Despite these efforts, it has been difficult to establish self-sustaining institu- tions that can msnage road maintenance efficiently or use external resources effectively. 3.02 Experience bas provided no standard solutions to the problems of institutional performance. Without proven formulas, institutional develop- ment has bad to proceed by continuous, local experimentation. Experience has, however, identified constraints on improved performance and some general principles worth pursuing. o Constraint. 3.03 Three factors bave worked against the institutional development for road maintenance: T the nature and constitution of the typical road agency, marked by conflicting objectives and functions and a lack of Incentives. * The weak public pressure for better roads. * The inadequacy and unrellability of funding. 1161 o D4 Structure and functions of the road agency. World Bank efforts to Amprove the management of roads and road maintenance have focused on road authorities, which typically operate as public sector monopolies* These agencies are generally responsible for three functions: 1) planning the development and maintenanee of roads, 2) negotlating and ovierseeing road Vork by contractors, and 3) constructing and maintaining roads using Its own work force (force account). The division of work between contractors and force account varies by country. Contractors usually undertake new construction and large rehabilitation projects, while road agencies do routine maintenance wlth their own work forces. Resurfacing operations are sometimes shared, sometimes not. In some countries contractors also do routine maintenance, and In some the agency's work force also constructs new roads. 3.05 Contracted works for new construction and rehabilitation normally constitute the greatest part of agency expenditures. But direct execution of maintenance orks, accounts for the major share of road authority employment and for a disproportionate share of difflculties. - 22 - 3.06 Conflicts may arise when a state organization combines zhe planning and control of construction and maintenance operations with the direct execution of works. Internal control is blunted and efficiency prejudiced when direct execution receives too much emphasis, if only because of the political significance of the large labor force employed. Already weak incentives for efficiency are undermined by the constraints Inherent in the public service. Civil service rules circumscribe managerial decisions. Deficient salary scales and promotion systems make it difficult to retain competent staff. And employment objectives frequently smother the technical work of the authority, distorting Its decisions (Box 3.1). 3.07 Weak vublic pressure. Because road agencies do not operate road transport services, they do not suffer directly from bad maintenance. Nor do they normally sell services to road users in a coupetitive market. They thus do not bear the full cost of neglected maitenance. Nor are they subject to market pressure, as railways are. Also absent is the pressure from direct dealings with the public -- pressure felt by public health, education, and transport services, for example. Truckers rarely experience enough competition from other transport modes to make them complain about the high cost of operating vehicles on bad rosds, and car owners and businesses are too dispersed to form effective pressure groups for better roads. Furthermore, the effects of the road authority's neglect are unlikely to be perceived before the problem has become acute. Without pressure groups to prod rhem, politicians and addinistrators have little concern for the roads, and the road agency baomes a facility for political patronage and unemployment relief. 3.08 Inae@ute budgetary provision. Without political support or market pressure, the budgets for road maintenance are often unrelated to known requirements. Because much of the budget tends to be committed to the wages of an unnecessarily large labor force, the disposable portion fluctuates much more than the total when budgets are cut or prices rise. A small cut can bring much of the agency's work to a standstill because of its inability to buy fuel or spare parts. Moreover, even when funds have been appropriated for maintenance, political or private interests put pressure on road authorities (or their financial sources) to divert funds to other pur- pose. Such interference, often one of the biggest wenagerial challenges, can be controlled only In the political arena. The Search for Solutonm 3.09 Accountability. A striking absence of accountability is a common cause of the Instltutional failure of road agencies. Planning, supervision, and execution are normally combined In the same orgSaniation. load agencies seldom link the allocation of funds to explicit physical plans ad rarely perform post-project evaluatlons. To Improve accountability, plannAng and supervision must be separated from excution of works, and procedures must be Introduced to enure the evaluation of project results. One way of doing thls is for the road agency to award maintenance contracts through competi- tive bids, and then to supervise and evaluate the work closely. Raieing the visibillty of maintenance objectives and performance Is equally tiportant. - 23 - Io 3.1 %MeM#tffM Lu i Ieuowrc laaame INob The effects of the general budgetaty squesse on operating fund (for ful, spare parts, anW bitumens, for example) has been eacorbated In Kenya by two parallel developrents affecting the toads Department. One Is the decision to integrate casual labor with the persanent work force. The other ts the Implementation of the Distriet locus Programs, a broad-based effort to decentralize governmest fuctions. Until 1982 the (central) toads Departmeat employed an average of 19000 permaneat staff a about 9,500 casual workers. Wages and emoluments amounted to 6.7 million Kenyan pouds (IL), and operating funds totaled XL 9.7 million, whicb was already Inadequate. Three years later, however, the situation bad deteriorated dramatically because of Increases ia the permaent work force and decentralisation of road agency functions. The toads Department's casual workers (and some persomnnl from other minor departmet) had been Integrated wlth the permanent work force at th district level, and the number of persoanel bad rise fro" 1,000 to 14,600. Of the KL 10.5 million allocated from the central budget to the districts, more than 90 percent went for personel-related expenditures, leoving little for operatlg expenses. MeanWhIle, funds for operatlons that were centrally controlled (Including ost equipmt-baed operations) were down to XL 6.3 million. I~... .... - 24 - Unless this is achleved, the chronic underfunding of maintenance In developing countries will be difficult to reverse. 3.10 Also essential for an agency's accountability is monitoring by an Independent authority. Performance audits should relate financial flows and physical performance ndicators to the state of the roads. An information system for collecting such data - needed for planning and for monitoring performance - Is therefore an essential lngredlent of good policy. The data can also be used to make an independent and public assessment of the agency's pevformance. In the United Kligdom, for example, county road authorities have to publlsh annual reports showing thelr costs, comparing them witb private contractors' charges. [91 3.11 Public awareness. Wider public awareness, Including that of potential nterest groups (contractors, exporters, transport enterprises) is isportant for shaping policy and obilislg support for programs to restore and maintain roads. The Swedlsh National Road Admnlaitratlon, for example, carries out periodic surveys (using questlonnaires) to gauge how the public perceive its work. It then use the results of these survey to redirect Its policies and operations. It also uses the results to Influence the Parliament and the Government to ensure idequate bigbway funding. The public information program of professional societies and trade associations - uch as the American Society of Civil Engineers, the American Automobile Association, and the National Asphalt Association - have also helped to arouse public rionceru about the state of roads In the United States. 171 In France the national and regional associations of contractors (YEddration National. des Travaux Publics and PFddration Rgisoale des Travaux Publics) perform a slimlar role, lobbying for Increased public spendlig on roads. And in Japan, the Road Association, an lnterprofessional body with private and public representation, promotes public Interest, both technical and geanral, about the needs and condition of the country's roads and acts as an effective lobby at all levels of government. Ileints of werm 3.12 Nanagement system that work in developed countries have not always improved the operation of road maitenance Institutions In developing countrles. And when they do, It is not clear wby. Reforms will thus have to be experlmetal, starting with an understanding of the Institutional environment and focusing an accountabillty. 3.13 Decentralization or functioal separat on. Decentralislng the organisations for road maintenace has oftee bn advocated as one of the first steps In reform. The presumption is that it would Improve responsive- ness to local needs and reduce the management difficulties Inherent In wide geographical coverage. One contrary argument Is that tecohical and manage- rial staff would be spread thin. Anotber Is that the increased political sensitivity could dlvert maintenancef unds to more popular (or more conspl- cuous) but less worthwhile construction activitles. It could also lead to the emloyment of redundant local staff - as In Keny and Honduras. The Bank's review of experience Is not conclusive on this point, showing little - 25 - correlatfAon between success In road maintenance and the degree of centralisa- tion. That leaves pragmatic case-by-case experimentation as the only prescription. 3.14 More obvious ts the need for separating the execution of works from the plannlng and control functions, for that would help to Insulate planning and control from the pressures inherent in operating a works program. If * execution were turned over to private contractors (operating under competi- tive Incentives) there would be benefits from the greater freedom to procure and manage resources. Private contractors offer better prospects for developiag an efficient, lasting Institutional capacity for carrying out road maintenance services. They can also constitute an Informed group directly Interested In the adequacy of road maintenance budgets. Experience suggests that private competltve orgsaisations use and maintain machinery more effectively than do government agencies (Box 3.2). 3.15 A review of experience with the contracting of maintenance In nine countrles found that roads under contract are generally well maintained In seven of thm (Argentina, Braxil, Central African Republic, Ghana, Ken'ya, the United Kingdom, and Yugoslavia). 1151 In the other two (Colombia and Nigerla), the problems encountered in initial experiments are being evaluated to flad solutions. Evidence suggests that private contractors can perform work at lower cost than goverment organisations. In one case with fully comparable costs under similar conditios (Ponta Gross&, Brasll) maintenance by the permanent. work force cost 60 percent more than contract maintenance. Contractors have been attracted to maintenance opportunities, even In remote areas. Very often, small local contractors can operate in remote areas more cheaply than can the central road adsiaistration or larger contractors. 3.16 The desirability of contracting a major part of ongoing routine maintenance and resurfacing dbesrves wide consideration. Retaining a small government maintenoane capaclty can eas the transition, provide back-up for emergencies, and reduce the risk of replacing a public monopoly with a cartel of private Interests (as occurred In Nigeria and to a lesser extent in Brasdil). Japan's policy of relying on contractlg for road maintenance is reinforced by numerical limits on the labor force that government agencies mey hire, An Increase In the demand for civil works results automatically In resort to contractors. Even the field offices of the local public works administrations can let maintenance contracts of a limited, but far from neglibile value. Equipmest In public ownership iS kept at minimal levels, just enough to meet emergencies. Thus the Investment In goverent-owned workshops and space parts Inventories is negligible. The United Kingdom bas had promising initial results with a program (begun in 1981) that requires the agency's work force to compete with private contractors. (91 This competition stimulates pressure for efficiency on both sides and has provided an lncentive to assemble reliable cost data* 3.17 Contracting technluees and skills. The use of contractors can reduce the burden on the road authority, but It lncreases the need for super- visory managemet. Harnessing the profit motive to get the required work done well, on time, and at reasonable cost requires knowledge of contract - 26 - Box 3.2 Low USiU1 Ut Utilization In Vast Afric and Latin Akerica Road maintenance survey, by the World Bank in 1985 for West Africa and Latin America found vehicle and equipment utilisatiov rates far below what generally Is regarded as efficient (1,250 hours a year). In West Africa, the average annual hours of operation ranged from 420 for rollers to 840 for dump trucks. In Latis America, utilization averaged 750 hours for equipmeut and 800 for vehicles; crushang plants were utilized even lese, and asphalt finishers had the lowest rate - 420 hours a year. In West Africa, lack of spare parts and fuel appear to have been major caues of low utllisation. About 40 percent of the road authorities received only half the funds needed to purchase spore parts to keep their fleets fully operational. Fuel expenditures In nine countries In 1982 were only about one-third the amount needed for 1,000 hours of operation. The lack of spare parts hindered maintenance, and the shortage of fuel restricted operation. Low utilization is harder to explain In Latin Amerlca. With three exceptions, road authorities In the countries surveyed had enough funds for spare parts to operate the fleets 1,000 hours a year. About 40 percent of the agSences had less than two-thirds of the funds neded for fuel for 1,000 hours of operation a year. But the rest did sufficiently better to raise the reglonal average to 82 percent of the funds needed for fuel. Poor equipment mageet and, possibly, too few well qualified operators and sechanes may partly explain the low rates of utilization. -27- design. The contract should not provide uneconomic incentives (as In a cost-plus contract) or restrict the contractor's freedom to seek cost-saving methods and sources of supply. Where institutional capability Ls limited, consultants, speciallsts, or management-services contractors can help to develop management systems and contract instruments - and provide training for government staff and contractors. 3.18 If well-established doaestic contractors are not available, the trial Introduction of small contracts (say, for routine maintenance that is technically slimple and requires little investment) can reduce risks and help develop the capabilities of government and contractor personnel. All successful contracting schemes have Involved close coordination between the governAent and contractors in defiAnig and plamning the work. Such schemes have also changed the role, sources, and contractual modes of foreign techni- cal assistance. 3.19 Human resources. Better use can be made of human resources In several facets of road maintenance. Unskilled labor can be substituted for machinery - a cost-effective solution when wages are low, activities are properly organixed and managed, and suitable Incentives aro built into the system. [81 And many countries have had good results with the simple "lengthman" system. Under this system, people living alongside the road are responsible for maintaining It, with payment and continuing employment contingent on satisfactory performance. 3.20 Low and Inflexible civil service salary scales make It difficult for government agencies to retain competent managers, engineers, technicians, foremen, mechanics, and others with special skills. Sometimes Incentives can help retain the most productive staff. The roads authority in Ethiopia, after some trial and error, devised a bones scheme based on work-unit productivity. Agencies In several other countries use special allowances to encourage field supervision by middle-level managers. But despite these and similar schemes, staff turnover is likely to continue to be high. Road authorities may as well recognias that one of their useful functions is to serve as a&training and proving ground for personnel who will move on to the private sector. 3.21 The personnel management - or aismanagement - is the area with the most potential for improvement. Staff mest be motivated - and incul- cated with a bense of duty and accountability - with good performance recog- nized and rewarded. Recruitment and appointment to senior positions must be based on competence, not patronage or antiquated seniority-bound civil service regulations and policies. Weak performance is all too often blamed on inadequate training when It stems from lndequate, nonexistent, or unenforced personnel management policiess Training whose Importance cannot be emphasized enough must be an lntegral part of personnel managemnt. For such dispersed work activities as road maintenance, road extension services can be modelled on the training and visit system used In agriculture to provide an effective and fairly low-cost means of training maintenance staff. 145, 461 - 28 - 3.22 Technical assistance for road mlatenance has traditionllly focused o0 advisory services, maage_et systems, and tralAng for road authority personnel. It is lses coon for technica4 asslstance personnel to act In an operating, rather than advisory, capacity.4/ An a result, providers of teachical serv"aes have little exposure to rlik, few licentives for Improving performance, add only indirect responslbllity (If any) for measurable output. 3.23 Some scbemes have tried to establish perfomanee Incentives for suppliers of technical asslstaone by redefiniag their role and la"ving the move of the risk. But traditional suppliers of toeical assistance generally lack the finalcial wherewithal to accept sgniflcant risks. So It may be desirable to look to nAw sources of assistance, such as lternatloual civil engleering contractors. These companies are accustomed to risk-taking ventures and performance-related inceatives, and their staff have the requi- site qualifications and are accustomed to workiag together as anagement teams in developing countries. Aother optlon is a "twvinning arrangement between a developing country road agency and a partner listitution in a move advanced country. Such arrnge _ mts can facilitate regular exchanges of middle-lovel maagers and technical staff. They can also provide for the transfer of technology, procedures, and systems. (471 The Federal Righway Admianstration of the United States has collaborated in such arrangements with the highway authorlties of Argentina, Ethlopla, Jordan, Uberia, Nhpal, Philippines, and Turkey, among others (Box 3.3). [481 france, through its techncal services agencies, has also provided assistance to develop similar institutions In Algeria, Casroon, and Cote d'Ivoire. 11 Among the exceptions are the technical assistane programs of France and the United iangdom, which have provided technical staff to assum senior line positions In the bigbhwy agncies of such countries as Djibouti, Cote Ad'voire, Gambia, Malawi, and Nlger. The effectiveness of such technical assistance wlll be greatly enhbaned If the tecnical inalstries of the posaoring countries provide adequate backatopplg. - 29 - Box 3.3 Sinnin od OS eand ishniym muiuatim In 1947 the United States laltiated a cooperative program to help Turkey develop a highway orgamosation capable of constructing and main- taining an expanding road system. The US Public loads Adinulitration (PRA) was entrusted with this task. The first phase of the assistance to the Turkish highway adainistration consisted of the PBA staff's assess- ment of highway developmaet In Turkey and a formal agreement with the Turkish government covering the objectives and arrangements for imple- menting the program. PIRA's staff Initially provided the core organization for the work. Specialists were added to the organixatioal units as required, but as soon as Turkish personnel ould take over an organiSational unit, that unit was dropped from the technical assistance program. The first division to be transferred to ec¢luslve Turklsb administration was the. Planning and Programming Diviston. When the techidcal cooperation program began in Turkey, local counterparts could not be found for many of the technacal divisions, such as Materials or Survey and DesiSgn To fill this gap, training programs were organized with the better graduates becoming instructors for new classes. While learniag on-the-job, the Turkish personnel gained further practical experisece as they built and maintained the highway system. And when recruiting satlsfactory equlpment operators proved difficult, the Turkish Army supplied me and officers for the program. The last US advisor left Turkey In 1958. By thea, the Turkish General Directorate of Highways had evolved Into a first-class highway organixation capable of maintaining 27,000 miles of all-weather national and provincial highways most of Which were Improved or constructed during the period of the US teechncal cooperation. Source: America's Hzhways 1776-1976, US Department of Transportation, - ederal Sighway Administration, Washington, D.C. (1976). [481 CIAfURt 4: IWCIMt 4.01 Pew developing countries have spent wisely or enough on road maintenance. Not all roads have to be maintained in the best condition to maximize economic returnl, but heavlly use road should not be allowed to deteriorate - especially wben a fairly small, properly timed expenditure for maintenance would make reconstruction unnecessary. The costs for restoring deteriorated and rapidly deteriorating roads are substantially more than most developing countries have been spending. But far greater sums are at stake If the situation Is not handled better over the next decade. Mm LagEgte Pietwe 4.02 The financial requirements of the road deterioration problem have two parts. One is the cost of economically warranted restoration (rehabili- tation and reconstruction) of the roads that have been allowed to deteriorate to poor condition. The other is the future annual cost of economically warranted maintenance of the whole network. 4.03 The cost of restoration. The cost of economically warranted. restoration of degra4ed main roads for 85 developing countries Is about $43 billion (Table 4.1).1/ Even witobut a claim to accuracy, the number provides an order oCragultude. Bridges and large tertiary and lower-order networks were not included, but a speculative estimate for such components is about $15-25 bllllon. Under limted budgets, only some of these components would qualify as economically warranted restoration. The estimates also exclude 8bout $3 billion for some rehabilitation of very low-volume main roads in the poorost countries of Sub-Saharan Africa. Full restoration of these roadt would not be economically wcrranted, although maintenance should be at higher standards than at present. 1/ The estimation procedures are described by Smith and Barral. 1351 A few countries tn each region account for a large proportion of the region's restoration needs: Eastern Africa: Madagascar and Zambia together, 402 Western Aftlca: Nigeria and Ghana together, S8X East Asia: China, 562 South Asia: India, 64Z Europe, Middle East, and North Africa: Turkey, Yugoslavia, and Algeria together, 562 Latin America and Caribbeans Brasil and Argentina together, 63X. - 31 - Table 4.1: Total Expeaditures Required to Moet the Estimated Restoration lacklog In 1984 (US$ billion; 1986 prices) Rehabilitation Total as a and reconstruction Total Percentage of Gup Region Paved Uipaved medlan Range Eastern and Southern Africa 1.4 0.8 2.2 3.32 0 - 15.5X Western Africa 1.9 0.9 2.8 3.22 1.2 - 36.9S last Asia and Pacific 7.5 1.6 9.3 1.62 0*2 - 2*5Z South AsiA 7.7 0.9 8.6 3.5X 2.5 - 12.72 Europe, Middle Rest, and North Africa 8.4 0.9 9.3 2.62 0.4 - 7.42 Latin American and Caribbean 7.9 3.1 11.0 2.02 0.4 - 22.86 Total 8 L (percentage of total) (81) (19) (100) Tle 4*.2: Annual Maintenance Expenditures Required to Preveat Deterioration, 1986-90 tugs biliiona In 1986 prices) Total as a Region Routine hlsurf acAx Total ,ercetae of oNP ' maitenace Paved Unpaved median Range Eastern and Southern Africa 0.2 0.1 0.1 0.4 0.72 0.2 - 1.82 Western Africa 0.1 0.1 0.1 0.3 1.02 0.1 - 3.52 East Asia and Pacific 0.3 0.8 0.2 1.3 0.22 0.1 - 0.42 South Aala 0.1 0.3 0.1 0.5 0.32 0*2 - 1.12 Ekrope, Nlddle BEat, and North Africa 0.2 0.5 0.1 0.8 0.32 0.1 - 0.6Z' tatin America and Caribbean 0.5 0.6 0.3 1.4 0.42 0.1 - 2.82 Total m4. =4%07 (percentage o! total) (30) (51) (19) (100) - 32 - 4.04 Future costs. Strongly affecting future costs will be the money spent (and effectlvely used) In the next few years to prevent the deteriora- tion of more roads Into poor condition. A remarkably high proportion of the paved network (more than 40 percent In 1984) already Is or will soon be in only fair condition, requiring major maintenance to prevent structural failure. Met promptly, resurfacing and routine maintenance needs should amount to about $4.7 billion a year (Table 4.2) over the next few years and then taper off soaewhat after the bulg of roads In fair condition passes through. If these needs are not met ia time, however, the costs willl multiply rapidly. An example of Inadequate spending is the $43 billion worth of restoration needed aow because of not doing some $12 billion of preventive maintenance over a decade or so. 4.05 Required allocations. The funds needed to prevent further deterioration and to clear the $43 billion rehabilitation backlog (in either five or 10 years) are summarlzed In Table 4.3. If all countries met the five-year target, the cost would be about $13 billion a year ($4.8 billion for maintenance and $8.6 billion for restoration). For the 10-year target, the cost would be about $9 billion a year ($4.8 billion for maintenaice kad $4.3 billion for restoration). The foreign exebage requirements would be $4 billlon for the five-year target mn $6 billion for the 10-year target. 4.06 These estimates are less than, economically optilal because they e-mphasize holding down road agency costs at the expense of higher user costs (especially In the 10-year program). That emphasis Implies gltvng priority to saving roads In fair condition to minimize future restotation costs, even if the social return would be higher from reconstructing heavily traveled roads In poor condition ad thereby saving user costs. 4.07 Costs could be reduced by Improving the productivity of maiute- nauce, but It Is not clear whether such galns are attainable In the short run. There is scope for Improving the productivity of some elements of maintenance operation - labor, equipment, and procureent and supply. But such gains are likely to be disrupted If the lastitutional structure is defective. Even If supported by am Infusion of technical assistance, such gains are difficult to ustain without the commitment and dedication of local staff and maasgers of road authorltles. Productlvlty gains are likely, therefore, to account for only a aall reduction In the short-tern bill. 4.08 Total road expenditures, includlng constructiou, In the 85 countries are estimated at $10-13 billion for 1984, of which restoration and maintenance may have accounted for about 50 percent. With $4.8 billion a year requlred for future maintenance, little would remain for catching up on the rehabilitation backlog. In may countrie the restoration and mainte- nance budget Is too small even to stabilize the road networks In thder present condition, let alone rebabilitate thm. - 33 - Tble 4.3 Annual Financing Requiremeants for Restoration and Msaintenae of Roads (US$ billion In 1986 prlces) Total cost Forelgn exchanae 10-year 5-year 10-year 5-year target target target target Eastern and Southern Africa 0.6 0.6 0.3 0.4 Western Africa 0.6 0.9 0.3 0.5 last Asia and Pacific 2.2 3.1 0.9 1.3 South Asia 1.4 2.3 0.6 0.9 Europe, Middle East, and North Africa 1.8 2.7 0.8 1.3 Latin America and Caribbean 2.5 3.6 1.2 1.7 Total 9.1 13.4 4.1 6.1 CJqnori Differences 4.09 There are wide differences among countries In the financial requirements for road maintenance and restoration, the ability to marshall and allocate the needed resources, and the capacity to use additional resources effectively. 4.10 Financial constraints. The financial saverity of a country's road restoration requirements can be gauged In two ways. One Is to estimate the lmprovement In current funding for maintenance tat would result from reallo- cations of the overall sector budget. The otber Is to look at the rate of growth of real GNP per capita during 1975-85. The first criterion shows what could be dons with current road funds. 'me second glves a crude lndlcation of the potential for licreaslng those funds. Altbough any Indicator of the potential for lcresring funds it open to questlon, the growth of Keal Gan per capita is believed to be a valid lndicator for most countries.-/ 2/ There are bound to be exceptions: the oil-producing countries ad others linked-to them economically had unuually high growth rates during 1975-85 but were set back by the recent drop In oil prices and are now facing adjustment problems. - 34 - 4.11 Current road budgets (from domestic sources and, in some cases, external assistance) were placed in three broad classes according to their adequacy for road network restoration: * Sufficient financing capacity. Current road funding, if reallocated, would be sufficient for adequate maintenance of the network, for complete restoration within five years and for new construction amounting to at least 20 percent of the total. M Noderate to marginal finacing capacity. Current funding would have to be increased by up to 50 percent and new construction held to 20 percent of the new total to carry on adequate maintenance and complete the restoration in 10 years.3/ o Insufficient finacing capacity. Even if current funding were increased by 50 percent and new construction were held to 20 percent of the new total, funds would be insufficient to complete the restoration in 10 years. 4.12 These criteria were applied to 61 countries for which some expendi- ture data are available (Table 4.4). More than half the countries with insufficient budgets to finance their road requirements (11 of 16) had negative rates of real CEP growth. Of these, nine are in Sub-Saharan Africa, where extensive road building and paving in recent decades set the stage for today's rapidly worsening maintenance situation. These countries have the greatest need for increasing their road budgets and the poorest prospects for doing so. Along with their neighbors in the table, they constitute the hard core of the road maintenance problem in the developing world. 4.13 A somwhat differen- problem affects China and several countries in South Asia: Inldia, Pakistan, Burma, and Bangladesh. In general, many of their roads were built long ago to standards of strength and geometry that are totally inappropriate for today's traffic. Although these roads, too, have deteriorated extensively, my of them should not be rehabilitated. Instead, they should be replaced by roads of a capacity adequate for present and expected requiremts. 4.14 eallocation of funds. LAving aside the special needs of coun- tries with obsolete networks, the reallocation of funds from new construction to rehabilitation and maintenance will be necessary for almost all the other countries in Table 4.4 unless they can increase the road budget signifi- cantly. For many,- the reallocations will have to be drastic. Even countries with "sufficient" financing capacity will havo to shift funds from one category to another - which In many countries is difficult because of the budgeting procoos. Sme have separate national budgets for capital and current expenditures, with construction in one and rehabilltatlon and 3/ There is a widt definitional gap between the first two categories. None of the countries fell within that gap, indicating a clear distinction between the two groups. - 35 - Table 4.4 Road Restoration and Maintenance: Financing Capacity in 61 Developing Countries Annual growth of real GNP Capacity to finance required restoration and per capita maintenance based on recent higbway budgets a/ (1975-85) Sufficient Morginal insuffictent *Korea *Botswana *Yen, PDR Over 2.5 *Yeman Aa Egypt Burma *Csmeroon *Paraguay Pakistan *iAsotho *Thailand Sri Lanka *ronesSA TUDisia Algeria Yugoslavia ungary *Nepal India *Burknia Bagladesh *Colombla bonduras Morocco *3enin OZ - 2.5 Turkey 8cuador Brasil *'enya *Iand.a *Malawi Chile Mali *Burundi Philippines Portugal Guinea *aXaan Dominican Republic *Papua New Guinea *Swasiland Sierra Leone *Gabia *Nlger *Costa Rica 'Mauritasia Li1beria Less than 02 *Cote d'Ivoire Uruguay Zambla Zaire *Nigeria ZiLbabwe Tanzania Jamaiea *Central African *Senegal Bolivia Republic Madagascar *Togo Argentina Note: Countries are listed In desceAndLg order of annual rate of growth of real GNP per caplta durIng 1975-85. a/ The recent national highway budgets (1981-85) used for this analysis inClude, in certain cases, external borrowings or grants. Tbe assessmet of financial capacity is somewhat subjective as highway budgets fluctuate significantly over tLme in many countries. See test (para. 4.12) and 1351. * Countries with recently ezpanded or improved networks. - 36 - maltenance-in the other. And In many countries, road construction is part of the genesral development plan - controlled by a special office or ministry, not the ministry or bureau responsible for maintenance. In such a system, reallocating road funds from construction to rehabilitation and maintenance is almost impossible once the budget Is drawn up. (A change in allocations would require a top-level decision before budgets are submitted. Such a decision would probably encounter strong resistance from the entity whose construction funds were to be reduced.) 4.15 Reclassifying the budget could ease the consequences of separating construction from rehabilitation and maintenance. For example, resurfacing and rehabilitation of worn-out roads could reasonably be classed as capital rather than recurrent expenditures. Capital project planners might then pay more attention to the need for maintaining older roads and for assessing the future maintenance requirments of new construction projects. The budgetary constraints on rehabilitation would also be less rigid. Such a reclassifica- tion should also be useful in reorienting the programs of external aid agencies. Another possibility would be to integrate capital and current bud- gets and to select maintenance, rehabilitation, and new construction projects on the basis of their economic worth - taking into account future streams, of maintenance requirements and road user savings in each case. 4.16 Absorptiva c!MCIty. Money is not enough, because expanding the capacity to use funds effectively can be very difficult. Most of the coun- tries considered here will have to expand their rebabilitation and maiate- nance operations, whether by shifting resources from conscruction or by other means. Road maiatenance agencles will have to handle maintenance efficiently and on a much larger scale than before. All this is likely to require a tremendous expansion of the mantenance capabliUties of the government, or private contracting firms, or both. Machinery will have to be acquired, operators and mechanics will have to be trained. And above all the managers and staff of the road agency wlll have to learn how to plan, manage, and supervise maintenance work efficiently on a larger scale. 4.17 3xppnsion may be especially troublesome for countries with recently expanded networks of paved roads. Many of these countries have little experience with maintenance. Nor are they prepared for timely and adequate maintenance that will prevent the need later for far more expensive rehabili- tation. Slightly more than one-balf of those 30 countries Included in Table 4.4 that have largely new paved network are judged to be problematic in terms of their absorptive capacity. Such judgements, based on Bank experience, may be subjective but the overall balance is realistic. The worst combination of problems afflicts some Sub-Saharan countries with relatively new networks: insufficient financial capacity, negative growth rates, and problematic institutional capacity. Nmhellsn Dometic Resoures 4.18 Funds reallocated from new construction and gains in efficiency will seldom be sufficient to finance the swelling costs of restoration and saintenance. A comon method for Increasing revenues to finance larger road - 37 - expenditures Is to raise road user taxes and other charges. But In many countries, revenues from road users already exceed (non-urban) road expenditures by a wide margil; so wide that urban road expenditures could not possibly account for It. / Taxes on vehicles and supplies are a common and convenient source of getnral revenue - often an essential source where the capacity to tax is limited. But high road user taxes may distort transport costs and lead to economically lnefficient business decisions about location, production, and Investment. When distortions exist, taxes should be ralsed only after their structure is adjusted to reduce distortions. For example, where heavy vehicles pay less than the cost of the damage they inflict on roads, a tax Increase should be scructured not only to raise more revenue but also to encourage the use of multiple axles to reduce road damage. To the extent that additional levies on road users are translated Into better roads, they will reduce (not increase) road user costs. 4.19 Enrgarkina revenues. Practically all the arguments of fiscal economics go against earmarking taxes for such specific purposes as road maintenance. Earmarking has nevertheless surfaced In a variety of develop- ing and developed countries and It should not be ruled out from considera- tion In two cases. In one, where fiscal control is weak and revenus alloca- tion and disbursement are subject to seepage or long delay - earmarking can assure that the government's decision to maintain and rehabilitate roads is translated Into practice. Short-circuiting the budget may then ake for a more rational use of resources. In the other case - where a special tax is to be levied (or a rate raised) to finance road rehabilitation - the principles of benefit taxation apply. The taxpayers, within a well-defined clientele, know what they should expect to get in return. Such a targeted tax, like the earmarking scheme, will have a limited duration. 4.20 In either case, earmarking can be recommended only If the fiscal Integrity of the receiving authority (say, a road fund) and its determi- nation to apply the funds according to economic priorities are assured. A frequent problem Is that money earmarked for the road fund does not find its way there or is used Inefficiently or for purposes other than road mainte- nance. One of the few reasonably successful schemes l developing countries emerged only after much trial and error In the Central African Republic (Box 4.1). 4/ Because expenditure estimates cover only nouwrban roads, It would be desirable to separate revenues derived from urban road users and tho"s from nonurban users. But this Is not possible because few taxes and fees are tied to the locations where the vehicles are used. (See Statistical Table C). m386 3.z 4. 1 es. :riemco vith Ubagkibe ld Numb i tb- Cantral Aricaa jejimlic In the 1960* and 19709 the Central African Republic failed twice in earmarking funds for roads. The first road fund was subject to so much discretiou that the eararking had no practical effect and the fund was abandond after three years. A second fund, set up In 1970 ia connection with the World Bank's First Road Project, was designed to avold the earlier shortcomings. But the amounts earmarked were lnade- quate, payments were erratic, and more mmry had to be sought through budgetary appropriations. A third road fund, inatituted In 1981 iL connection with the fourth Uighway Project, Is still functioning. As a public establishment wlth financial autonomy, the fund bas the sole function of flancing road maintenance. Revenues come from an iacreased fuel tax, which the govern- sent adjusts yearly to maintain the fund's capacity to do its work. The fund is under a ainisterial aget coittee, on vwhch road users are represented through the Chamber of Commerce. This third road fund has also had some problems, mona fatal. At one time, it had difficulty collecting the revenues. At another time, the government forced the fund to pay for a project outsild Its mandate. Despite these ad other problem, the fund has improved road maintenane by raising the level and lncreaing the regularity of fundlng. - 39 - lutrml in~oi ad Assistanc 4.21 The sheer esxe of tbe task now facing many developing countries Implies massive demands for external financing. The foreign exchange compo- nents of the financlig require_wnts range from 30 percent for routine mainte- nance In midle-lncome countries to 70 percent for restoration of paved roads In low-income countries. In many cases, extornal flaicing will have to cover nore than the forelgS exchange component to preserve the Infrastruc- ture. And In the group of countries at the hard core of the problem, no solution is conceivable without external financings 4.22 Ixternsl financing will bave to be focused oD restoration and main- tenance if the countries' efforts are not to be divertes to lower priority Investments. In World Bank ledlAg. for highways, this focus has progressive- ly sharpened. Over the decade ending 198S, 53 percent of World laok lending for hlghways ($5 billion of $9.4 billion) was for rehabilltation, mainte- nance, and technical assistance largely relatet to maintenance. Soms multi- lateral and bilateral agencies, particularly those with a longer experience of providing assistance for blghway development, also raised the share of maintenance and rehabilitation In their flancing of roads. But most of thems, continued durlng 1981-84, to provide assistance for new construction, even to countries with mounting maintenance backlos. llateral aid tended to have a higher proportion ofn aw construction and lmproveent works than did mltilateral asaistas.ce (Table 4.5). 4.23 In gsearal, It makes sese to use external finance primarily for increasing a country's capital assets - an domestic finanee for covering curreat costs, such as those for maintenance. (ThIs is the tendency In most countries In 8ast AsLa and in Europe, the Middle last, and North Africa.) It alo makes sense to regard the willlngness and abillty of v country to pay for maintennce - an ongoifag cost of road use - as a test of good Internal managemente There naturally are exceptions, mot obviously when countries have suffered major calamities, when the structure of demand bas changed considerably, or when the policies of lenders are distorting factors. None of these exceptions will, however, justify leading for new construction if the roads are In serlous disrepair. If donors and leaders restrict financing. to the foreign exchange cost of projects, and oepeclally if they show a strong prference for aw construction, they tempt recipient countries to divert their limited funds to such projects - as leverage to lncrea the supply of foreign aid. Under these conditions, It Is not surprising that malatenace io aneglected, especially In the poorest countries wher foreign funding plays a big part In determining resource allocation. Malatenance work is then starved of the Imported fuel, bitumen, and spare parts needed for the efficient use of equipment ad labor. In short, the icentlves are blsed agalist preventive maiateaace, contributing to the premature deterio- ration of assets. 4.24 If the deterioration of roads in developlng countries is to balt, all major lending agencies will have to liberalis their pollcls on lendlng for mistenance They will also have to coordlante their policies In each -40 - Tble 4*.5 Eternal Assistance for iigbways, 1981-84 percentage Distribution by Major Compovints Source Amount new Const., Rehab., Main- Technical ($NoM) &Ibp____e___ Reconst. tenance AssistanceO/ IBRD/IDA 4,344 432 342 13O 102 ^I other Maltilateral 2,889 Of which: IADB, AfDB, BADUIA, IOD, Islam DB, OPEC Fund (1,892) 392 302 102 12 b/ Bllateral: DAC 1.529 Of which: France, Germany, and Japan (864) 622 252 82 52 Bilateral: Other 450 942 52 12 neg. aj Mostly for maintenance. b/ Self-standing teohnlcal assitance (TA); project-related TA not included. country - rather than pursuiog dlsparate and sometimes conflcltlg goals. important in establishing the foundation for this .coordination Is having the various led ers agree on a comprebensive road program with the recipient country. And to ensure fiscal discipline In the long run, all parties should scrutlinze the implications of such a program: the capital and current budget;, the relative prioritles of different parts of the program, and the size of the program in relation to resources available from domestic and foreign sources. That should sake It easier to coordinate assistance in ways that make the most effective use of resources. 4.25 Chile shows what such a coordinated program of hlghway sector lend- ing can achieve (Box 4.2). The requlrements will differ in each country, so the content and relative asic of laternal and external contrlbutions will also differ. For countries at the hard core of the road problem, the requirements in each dimenson of the problem (money, techaical skills, and administrative capability) are greater than can be met without assistance. For these countries, the need for concerted action with donors Is particu- larly great, since there is so little margin for vasteo - 41 - los 4.2 World Bak load Sector I IMg In Clile Until 1985 World lank lending for roads in Chile was mainly for individual construction projects. By 1985, however, the rapidly rising road maintenance requirements were already more than Chile could handle with its central funding and its decentralized institutional arrange- ments. seventy percent (56,000 kilometers) of the network was the responsibility of municipal governments, which lacked the technicall- managerial, and financial resources needed for the task. On advice from the Bank the central road agency (Vialidad) decided to resume gradually the responslbllty for local roads and to use private contractors wherever they were competLtive. Comprehensive sector planning was introduced, using the Bank's Highway Design and Maintenanee Model* The central road aintenance budget was to be increased from $78 million for 1966 to $123 million for 1988, with two-thirds of the invest- ment budget reallocated to rehabilitation and reconstruction ($44 million a year). To support these moves and the new 1986-88 Road Investment and Maintenance Program, the Bank provided a sector loan of $140 milllon (21 percent of the program). Disbursements from the loan are tied to work on maintenance to ensure timely execution of this high priority program. The Bank also participated In a parallel $400 million loan and worked with other leaders and suppliers to finance the rest of the program. WAP= Ss POLICT COCWP ios AnD 5.01 The conclusions and recommendations brought together in this chapter derive from three main considerations: the increasing rate of road deterioration, the Insulation of road authorities from the consequences of poor maintenance, and the very large costs and financial requirements involved. 5.02 The dynamics of road deterioration have much to do with the road maintenance problem. Paved roads do not deteriorate at a uniform rate. o During an Initial phase of several years, deterioration is minor but it acce- lerates rapidly thereafter. Casual observation of road surfaces gives little warning of the iminence of the critical phase, when road conditions begin to deteriorate rapidly. If maintenance is neglected, road users bear the brunt of the Increase in total transport costs (the sum of infrastructure and vehiele operating costs), since the share of agency costs In the total cost of transport is small. Although vehicle operating costs constitute the domi- nant share of total road transport costs, improvements in (uncongested) road conditions save less in vehicle operating costs than previously estimated.. Project analysts, therefore, have sometimes erred by crediting road invest- ments with greater benefits than are justified. Meanwhile, road authorities bave often erred by ignoring the effect of neglected maintenance on user costs. For road paving, moreover, the two errors do not cancel each other out. Instead, they reinforce oe another: paving Is done before It is warranted, and then the pavements are neglected. The pavements become more and more costly to repair, leadlig to further neglect and premature failure. 5.03 Good choices in road management depend on many factors: climate, input prices, expected traffic flows, vehicle types and axle loads, existing road characteristics and conditions, efficiency of work execution and mainte- nance, available roes"-rces and the opportunity cost of capital, and attitudes toward different rl ,. Because these factors vary from place to place, uni- versal presriptions could lead to many more mistakes than successes - thus making It necessary to come up with specific solutions for each country. Some general conclusions nevertheless emerge from the empirical research of the past decad. Road Gravel roads are more economical than paved roads where traffic volumes are low, the climate is not extreme, construction materials not especially scarce, and adequate maintenance can be expected. This is true even for traffic substantially higher than the previously assumed lImits for these roads. The uncer- tainty surrounding some of the assumptions on which a decision to pave has to be based allows a broad range of cost tradeoffs with the breakeven traffic volume for paving varying from under 100 vehicles a day to more than 400. It may, therefore, be reasonable to make decisions on grounds of prudence, taking into consideration local conditions and the reliability of future maintenance. -43- * Some paved rOids should, according to present knowledges have been left unpaved because they carry so little traffic. Rather than maintin such roads at nomal standardts, It may be economi- cal, when funds are ltited, to let them deteriorate and simply to use low-cost patching to keep them usable. * If a road carrying more than about 500 vehicles a day is to be paved or strengthened, and If axle loads are hard to control, the econmic savings from staged construction of the pavement are lIkely to be lees then the cost of premature pavement fail- ure. Tbus, countries that have difficulty enforcing load limts should build roads to higher initial standards - even though this normally means building fever roads. * Because of the nonlinear deterioration pattern of paved roads,- the resurfacing of nevr roads can be deferred with only a small pealty - but only if those roads are not too close to their critical age and not heavily traveled. If, however, the roads are somsehat older or traffic Is heavier, the deferral of resVr- facing could cause Irreversible brkow and require costly reconstruction. * in ranking the factors affecting mainteace cboices, traffic volume Is generally more important than road condition. So wben budgets are severely constrained, scross-the-board maintenance cutbacks may not be the answr. It may be better to maintain high-volume roads In fair or good condition and to reduce sub- stantially the matemaace of sawm low-volume roads. B lecause the deterioration of pavd roads is go Insidious, a road maintennce agency mut regularly monitor the condition of roads so that aian will not be delayed beyond the point at which costs rise steeply. Tbe agency must also monitor the volume and mix of traffic and axle loads to deteraine priorities for lnvestment and maintenance. An effective anagement infor- mation system that covers these conditions as well as the agency's equipment, supplies, ad personnel Is a basic require- ment for adequate planning and dt ployuent of resources. * For saixum economy, road maintenance policy should be coordina- ted with road design and constrwution planlng, and the life- cycle csts of the roads should be balanced against the operat- iUn costs of the vebicles using thes. Tbis coordinatlon requires a good data base and a capable staff uslng sund analy- tical teeciques. Appropriate models and information mnagement systems exist for such anlysce. Their use does not prevent errors that result fro bad data or incorrect traffic forecasets, but using the systm may belp avoid the errors of oversiplifi- cation. -44 Iatitu*tiiml Dvewlom"t Inadequate maatenance in developing countries has various causes, but institutlonal failure is the only explanation for Its extent. At the heart of this failure is the absence of accountability. All activities to strengthen institutions, enhance inceatives, and improve the internal workings of road agencies should be judged by their promise to increase account- ability. The road agency itself should be subject to an independent system of auditing and inspection. * The nature of road deterioration and the separation of road management from road use have sheltered road authorities from public pressures and market signals. Both user groups and highway adianistrations should stimulate public awareness and communicate their needs and problems to policymakers. To limit e potential for conflict between the planning and control function of the road authority and the work execution function, the two functions should be separated. h goT erent should minimse its direct role in works execution, even for routine a3ntenance, and transfer that responsibility to independent entities operating in a eompetitive environment characterized by managerial flexibility and efficiency Incentives. i n countries with a backlog of roads needing rebabilitation and In thoge with young roads pproaching the age when maintenance requirements multiply, adequate resources sbould be brought to bear on the maintenance problem before roads get worse. This focus often requires reallocating funds from new construetion to rehabilitation and malntenance. Ways of incressing the total roads budget also should be sought, includitg linreased user fees and taxes, possibly earmarked for rebabilitation and main- tenance. Por many countries, external finsacing will also be needed - without it the serviceable network will have to contract. * n many countries road allocations will have to be transferred from nw construction to maintenance unless the total can be increased. Where budgets for the two purposes re separate or where a single authority cannot reallocat funds within the total, road resurfacing and rehabilitation should be elassed as a capital expenditure. 5.04 Recomendtions for Bank policy must distinguish between principles that apply to all highway lending and princwles that apply to different - 45 - countries according to their capacity for flnancing estimated maintenauce needs and the capacity of their instltutlons. A review of information, backed by the experience of World Bank staff, suggests that countries can be placed in four maln classes according to: * The state of the road networks. * The financial requiremets for maintenace and rehabilltation and the country's ability to meet those requirements. * The possibillties for reallocatlg recurrent funds for roads without major disruptionb. T The institutional capacity to absorb more funding for mainte- nance and rehabilitation. 5.05 The four categories are reasonably exhaustive and cover the different case that Bank operational work is likely to encounter in the next few years (fbw 5.1). Bach category can be exemplified by countries for which available Information appears adequate. * Category I countrles have, for the most part, used resources cost-effectively for road investments and maintenance and now have a serviceable road network. Sustaining past trends will dopend on the ability of tbhse countries to adjust their road priorities to changing circustances and to maiutain their existing institutional capacity. Provided these conditions are met and barrlng unforseen events, countries in this category .should suffer no serious road transport constraints over the next five to 10 years. For them, the main emphasis of external assistance sbould be on lastitutional improvemeats, technology traosfer, and operational efficiency. a Category II countries have underfundod road maintenance, with the result that It has been iasdequate In quality or quantity, or both. They could, however, rectify past inadequacies over five to 10 years by reallocating resources to maintenance, mproving the officieny of their operations and managemeat, securing nore external aosistance, and, In some cases, increasing the road budget. * Category III countries bave backlogs of rehabilitation and maintenance so large relative to domestic funds, personnel, and technology that correctlve action will take subscantial tlim, effort, and.external assistance. Thls category Is subdivided Into countries in which preserving a young network is the primary need (1I1t) and countries with older networks needing massive rebabilitation (111B). * Category IV countries liclude Chlna and most of the lndian subcontinent. Their conon characteristic is an estensive and obsolete road network In need of modernization to meet the -46- lox 5.1 Diagmsti Framework ftor Uxteral Assiatamee Polly t. the load Sector CATUSOUT to In these eountries, the institutional capacity and past maintenance effort have generally been adequate. They now need to adjust policies and expenditures to meet emerging maintenance requirements, particularly for now networks. They also need to devote continuing attention to institutional and technol*glcal improvements, and efficient operations. Examples: Chile, Korea, Niger, Malawi, Yemen Arab Republic. CAUGOB. II. In these countries, the funding for maintenance has been insufficient. They urgently need to reallocate domestic and external resources within and to road sector. They also need to devote greater emphasis to policy reform to expand institutional capacity. Examples: Kenya, 8razil, Nigeria, Indonesia, Yugoslavia. CAIEGN= ItIt In these countries, the institutional capacity has been grossly inadequate. They urgently need to mobilixe substantial external financial and institutional assistance. They must give a high priority to the coordination of aid programs and emphasize instLtutional reform. A. This subcategory of countries must give priority to resurfacing and strengthening paved and unpaved roads to preserve relatively new networks and to building instituttonal capacity. Examples: Mali, Benin, Nepal, Burundi, Liberia, Senegal. B. This subcategory of countries mast give priority to restoration of aging infrastructure and to building the institutional capacity to aid economic recovery. Examples: Laos, Chana, Bolivia, Tanzania, Madagascar, Sierra Leone. CAMP= IV*. In these countries, the networks and maintenance technology are obsolete. They need to mobillse domestic and external resources for modernization. They also need to place major emphasis on technology transfer, Institutional improvements, and the development of skills. Examples: China, India, Pakistan, Bangladesh. - 47 - burgeoning demnd for road transport. With few exceptions, these countries have not neglected maintenance. But their aged road networks and outmoded maintenance technology seriously constrain good maintenance practice. These countries need to mobilise substantial domestic and external financial resources to meet the growing needs of their road networks. 5.05 A common requirement across all four categories, and one of major importance to Bank policy, is the reallocation of domestic and external, resources to balance new construction with maintenance, rehabilitation, and betterment. Reallocation will vary according to each country's circustances within the broad limits for each category. In most countries, reallocating the road budget will leave fewer resources for new construction, so that new investment priorities will have to be established. In some countries, more- over, no amount of reallocation will suffice - nor will the economic situa- tion permit additions to the road budget. In such circustances, the net- workt's quality and extent will have to be reduced. If such a contraction appears Inescapable, the Bank's aalytic work will focus on establishing priorities for essential maintenance and restoration. 5.06 The Bank's sector and economic work in transport will refine the assigme t of countries to the four categories. Moreover, the analysis of a country's road sector will be a key input to Bank country assistance strate- gies for that sector, particularly for countries having the most critical problems. Because restoration and maintenance needs vastly exceed the resources available for this purpose in most countries, public investment and expenditure reviews will recomend the division of road expenditures betveen new construction, restoration, and mintenance. These reviews will also evaluate the financial and institutional capacity for coping with the future mintenance needs of new road investments and propose cost-effective Investment and maintenance strategies. 5.07 Bank lending for roads will be conditional on each country's achievement of an acceptable distribution of expenditures among major purposes: new construction, restoration, and maintenance. For countries in Category III, lending will normally be confined to road saintenance and restoration - with the condition that the country also applies its entire road budget, including external assistance to maintenance and restoration. Exceptions to this condition, other than minor ones, will be made where development roads are needed to exploit agricultural and mineral resources in the context of Identifiable investment projects. F Por countries in Category SI, lending will be conditional on periodic agreement between the Goverment and the Bank, about the division of funds among new construction, restoration, and maintenance. Funds will be comitted or releaed in tranches, conditional on reaching the Interim obectives of agreed-on programse - 48 - In Bank appraisals of road construction and betterment the probability of a road being maintained to acceptable standards will enter explicitly into the calculation of a project's expected net benefits, and reflect the country's past road mintnance record. 5.08 Because so much capital Is at risk and the economic outcomes are so sensitive to country differences in networks, trafflc, and other circum- stances, the Bank will require that road agencies have an adequate road ngeent system - or a phased, monitorable program for establishing one. And to improve efficiency and reduce the potential for conflict between the plamning and execution functions of the road authority, the Bank will encourage maintenance work by (private or public) entities outside the road authority - entities operating according to commercial principles and preferably in a competitive environment. In addltion, the Bank will support schemes to Increase awareness of the need for management systems and for adequate resources for rohabllitation and maintenance. These schemes will focus on efforts to educate goveroment officials and to make the public ware of the consequences of neglecting mintenance. Such efforts will. Include regular publication of the data from moitoring road conditions. 5.09 The Bank will also assess progress in the educatlon and public awareness components of Bank projects and In the transfer of maintenance work to entities outside the road agency. In addition, the Bank will periodically review the four-category classification of countries and make appropriate adjustments based on new evidence or the findings of its operational staff. Actla by the 5.10 In financing highways, external agencies should emphasie rehabil- Itation and maintenance In countries La Categories TI and III and moderni- zation In countries ln Category IV. They sbould limit thelr flnancing of new constructLon largely to countrLes ln Categories I and IV - and then only in integrated programs for constructlon, malntenance, rehabilitation, and modernization. In addition, all external agencles involved in road sector activLties ln any country In Category II or III should requlre that the recipient government establisb Lnternal admeinstrative mechanLsms to coordLuate and monitor external assistance programs* Consultative group or round table _mtings can offectLvely ald governments used to achieve this objectlve. 5.11 The data base needed for understandLng the factors that influence road deterioration and the effects of different maintenance processes should be substantially strengthened and refined. Road research institutions and bighway authoritles in developed and developLig countries should continue research on the multl-dimenslonal isoues of road deterLoration - technical, lnstLtutional and financial - and should support lnternational exchanges of data, Laformation, technology, and systems. In additLon, external agencies sbould provide technical and financlal support for the compilation of inter- nattonal statLitLcs on road conditlons. Such a compilation is essential if - 49 - worldvide trends are to be detected and reliable Judgments made about the relative performance of countries. 2he United Nations Statistical Office - or an organization such as the Permanent International Association of Road Cougresses (PIARC) or the International Road Federation (IV) - could assume responsibility for compiling these statistics. Also recowmmeded is the annual publication of statistics on the sources and uses of external aid for road development. temnxss:umm:nsssm: mu ~.:sssssssu/n x aui. ssssssssussoa was s : an MD mow: muNAm - ou WSu ai :: main : in : iil tm 3184 umwmmis m .. m : mum: M : : wA:1m.unIr:UM I: U11_JUM : Us:: ~~~: u .:m t: : : aw :p..sl :t tUm * : urru wan.: £ : n:n :urrum* : s:ii : mam :va. :a :un : is :e:: : :: m:. :1111111. MINO: t: W m: a : : : : a: .. . . .a.a. . . . .a.a.I* a . *a. * * . a*l.l. .. ::U 1111 ICZM is # 416 $,1: 11:20I: Sit: 4.1: .1 lN .t: U.: LI: II.: 110-1 '.11.A I.t: 0.1: I.: .t o.: t"l It.J: I.t It.t :IU N ::ha . su : 's:It ,IN : 3s.3: a : M: 4.: Si : .3: :.1 I.s: 1.1: MA M : a. . : a.t: a..: LI: mw: a,m : a:: a.: :a :: :4118011ding Mn I MU 11# t* : M 1. S :a.a : M: L Lla SA: LI : L : a a: SA: 310 a, : * 1.1 3:*.1 :It.#: a.. * * . a . . . . . . . * . S I S S S I.a .. ::umum wau : *. :s.I:,:t .':tiu: no: *a.o:a.: uls .6 1 .e1.e::.s: e.a: 144.6 :it: i.;:e.a: a.e: Ii..: 1.1: 4,u0: ee,m:as.t: s.e:a6.o;a.u:: ::NM An& i : I : ,wa : a..:: : o :1 4.: 1o.1: : :: ..I: st .: :ta. : i.1 : : a.: a. : 0.1: 205,: 41m1: to: a.s: a,.:IL. :: ::1iS 111i :ai 6 :I'ml: tS.I: M: In: t.4 11113: MI1.1: tl .|:1. *.s: U. :66t.11 l: 0.| 1.4: SIA.: LI: ,":. 0,: 1.: 1.S: IIIJ MA.§:: .. . . . . . . * . . . . . . ....... . . . . . . . . . ... ....... a . . . . . a . . . . . a. ::Auv M M C: 14 $j,tS :3.14 0: : IM. 3.:11.. : 1. 1 .1 C . 403.1:.4 5.4 0.2 4.8 31. .6: 2 e 5.4 .: U.4 :1. :: ::u.tish : : tW a.:w : : s: : .5 5: . 61:5 . : 2.0 : :.: :62.: 6.:0.: .:' te.: s.: z m: e.: : .: :21.:3?:: ::Ul U a: : : : : : : : : : : : : : : : : : : a : : ,.,.CAW=: : : : Is:olU- :" III :. :ll. 18 3 :13 : .1 SIM M : :- I : :. Il-: :. 4: 0 : . 1: . : 4: ::1 ::aosZi :fd 11 :,n 0.:IIU .|nl %1|14|2 .:ltUX 3l4 .:.: tX X:n:ta s2.:.: ::umu auaa : : : :t3 11.t : : LI : 11i : HA: I .: : 6 : at.,: : 0.2: I.:: i.e : mo : ,M: U..: :. :. :: ::3lj : ': : " : : tn: a.o:1w : 1.t: g.: LI: 2.: :iA : e.i:o : :3 .1t: 1a: m: : .J: .: .1:: ::auauma : ii :.,se16:ua.e:t:23 : s.:asa.o:l4Jlg: a.s:asaes It 0.? t.1 .§: *.a: *.o: 226.:: 10.6? 30 : .om:.e: :l.a:ss.e:as.':: .. * . . . . . . a . . . . . . a . . . . a......... . . . . . oo ... .. . . . . . . . . . . . . . . . . . . . ........................................... 1: vwU in buadk f. ::h.adimoususla : : : : : : : t : : : : : 2 t t : : : : : : : : ::~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - 4 sotatU 1.3 ~~~~~~~~~~~~~~~~~. . . .... ....... _. .... __. _._............ _.... . ... .... _..._._...... _._ .. . ..... .. .. ... . _.. .* ...... ......... an':. #is ets'ts* #. ai se 'nut t s Ss rl VI- tn _ a a wuinu a ~ ~ ~~ ' a"' a*a l t a0 s O tlt V III i't1 1' I' Sli @t M% Su'" I'll m a 1 an I to Kg aa a gmt' It at aft a t Ill &1 'l §l rG1N1't 61' Al Vit ot'tl All, U 11 Is' vi t ' n N Ut 5a u a ful't is P M. Ct et I i's ss si11rs v': cI rt uni R' " nn o u as M am M tt U utt vi a ot u m st '1 1'1 iiil I't tNI l'l woo VSt rt rl' wt l lt 911 0' 116 I 4 11 I ml# a6 metam#| a a a clonl l *s lt Ill "Il Ill Ill 'll ollt PlA I's- ow u M# "In a it M It a a mo't am NW" a n M i,: ri Vt I' M'st 'l^l lS III - 'u Ul'l Ills W Mol all =a u a s 51 i tl a t16 55 aina a "ls r s.VIre alt res re sa art us aI- es " i alel "rna w on a Iit 11 rtt ati t US in i" vi a Mnl uI a' s I sr m al rs Ill 1'1 i'ti M's re 111 is K'1 i-ns e amI n aulU Ul na ft I l a 'In II sIll rt us's IIs VS Il mot a it- uo W aIIt W M a ci c m ti ua nags a M na uu'`s tI I-s ii nn Ui ra c, m' me' VS. Mii Ma u's sim us t is i ast e l "na I U I# a i sa Ins s1 Vils ter Ila Il l I' mls sari t-at aM' all mu m la a sa I O slt tantIn uP usiI s rl t Vii's ut's VI rt aIs s's nt' to- rnt Mt It$ _ a1m ae Hi sat's Isa IIes is at it lusll tltIl l ases its' s1 ci I'l WINl SKI#' ri- M Ms' It," u III I m 01,1-111 Is I a 0 O-t V-t mo' - I It tI VI t'l a at " at t- _ *1118 i ia t sisitas ' sritu a sto i a a i ii VIVOe re- sel' rat Ill u m's an Mie Mi ris a' i at s1 a n m it tI no is M a1 s i r lt r-s VS mot ta ss S n'it USe sl st I ns Oum a ml us a It Fss n I n . a1 is a see sit s-r or ults r- e* s-Is sele' Mns I.) IT- an $is l t U iss ua a ss a AU als ni a I a l r 1 1's n e lt Ill Vst lti Ol'lt %ti (l Su s us e t 111 tM " n Iet a t W a s a a r s us'a tI s -s ci mee'I'll ul moll i s s Vs M ll l 1it n t ueltein lii u n tti nis Ui lire in e1n Ot n si 51 it 'i5 in1 t0 in Us a: _T U MluSt nea sIt nla mns os Is t 1r III sme i -i ss mot 1al re la of l$ an M V4 a~~~~it isu's ses aas i a us's 1 si i uss i is rt s,ts lis s se a em a s a aa s i s a M's eV sirs ins m1, fa rs u's "el' l-.e as las We ai aM ^ ~ ~ ~~u siets't sin ti ts ii tie, uses is ii ee tus' gis': s-s1 sat in us t on is s n " a I at isa l l Va two- usl VStl M's Ol ets i sU u's "on in It a is OWN sMe'O VI re n* l srts s ' s ilut llO'l MSt 41la on M a a s a G It I Al 19 0 a lul Ill Ill Olt ml Isn * V: I'll l WS Olt 91c ont M' ",t "M ti Hi s U i s UI nS' ts n t' tu i 51 ii Us 't ii sl 515 WigS i't it-t us Itt Wi| _SI v s all is a Is stl u a In a; M t " s te i' s ll m i a m s-si s-s mutz usil al- sn M "In Inm M • i s tal 1110t1t "M us is s ff Vs ml au s-irst seE *U 'i Olt M t op 6 1- Mii iS esu asn m •t asa 'et si rsu i an a se Vs re re tws r- re res iniissl Ill. M mas ot mlli M st.uueusu I u a "I'llaua oa w rtn rn ow' rot 1-1 fusw wet, esl usis U`$ qM sIte iss n a s sit ts aleas 1 a n ma'l es re ss ia V lll it or" sis ri at ati irs osna i a ft a u'l tson eI a a Wo o t SIli al'ls I'll re in'W sari s-s-s o n We a-s su . u M Ms n ti a It a in Sa t-l tI'st 5l i's $,t- m't Su all on i 41i li MU I uM a t is inu anm a Me':l 5I a t me -se rsl rl ti s-l re I S nlw ea s's I's M Me "au am sM tt ii tt M'S a a-rnt a it a 's't I's Is - 'l m'1 C's s-l l rt s's s't s-l n1I mil "u- n w Ms U M I i e a t M ISus s snul & lII ea Itll. M's tll Ill i a us imAue u nsisim it s t tis a an ils al-ittu Su I.e srt rr i i hI M I 1-s am M Is a a mile a-n cs s at 4 as 0ImVS U s t S e 11 r . m's tI l ti3 ml' uu' re MU is als _i1m1 111 a nof ss's W n tn te M I ai asmt l Vlils u's 1se1 rU s-s mtl us ra M s se1 muen sm is) WN) IS)lIi) I) (U) I) 1 (U) I) l Is) la) am m ls u- : "I iu) am mu ml 11 I'm) Wes tese) Iii ---,g -#101--miam --*JJ5tlgs-4- --sc4oo-- am * * s..--- - -- ----s .------ -. --- --g ml umm l Vu tmsl A1u1u1a M= *4UIiiemUOman In unmilum "Ule u_n "Mt la ---- -- --- *l5 5lms------. -u -- ---. -*--- ------- ', -- :-s - InIu _ 31U u f_Am Ol :5 8 uisaUIsm *;qttmiq q 1t smiqen III giadgi PM abasing p.g Igq:m Itu tmuainu IiKis ttltt :tS P 1i'i@' n'll it'O'' 56 tS't UUUt P4 I ._.. . .. __ ........... _._ ... . . __. ...._ . _._._.._..........._....... . .................. ;~~~~ ma n ttins r ni n tt U io a rs m tr tus I s m: ue a:- sut as i:aui ,,, ~ ~ ~ 3u r Sl'li 836KS Itlt i-i K- sees s'it' e.s: us i ns ii 9 iS Stl 1.1 tt mu a N urn net: N .a o 4 Gt " | tlt't to U111 l ta|l-li t-t 9-9 I'l l s-t letT "s -l u "*l t on- w 4' u of N m 0 l I U tt n n its as.l. no rI i I n sst t-t rrt use:t: r?en e n u I- #t' n-na au an I It n IUst m1a ams amw It is it 5 Is I1 r In'll II 05 sIt at': WtI' I'- utI it ttt Uam in v n nfI , t e i s n it o ot I. la of ofu ust Is e,of u t aull un? uu'; taul ml Utul0 ii an 1 g° ZI'tt *t i i'lt fulllts t t s K SI n5 n3'! 0 eta tttt S tl" A 01 SIt Uhf i's 01 t-t MI" ' 3 tit i1 r 5"It ro talt Al Wu6 U= n 0W t ne n Islt I n n.11 U fl I I'I o t't sai I i $l l meati air lU uau ll sti wu an *un nt e s al': si, Ilut t a t: us Utl sos6- et r:1 a:S vs' u- - - l t'u ga's, no mtZ i' n- _tu an sels a sn I at I Ut ItOat eit it ml) - r ID c tWlt t-t ua -l itt ttt' 113, t u, al"I alal fi"m an U it a'ut 11 m I It e n s nI . t' t U l t-s li'll IIt $It flu n11 m Gh t' all aU n " Sa 13 n 11'5 st ae eutt us ts I't rs rue III ws ut e t-t I's el 't uutn It sent a all uh an ai K U tt l I It I I to: tt u a tS nt sI rIes I's t uI nt r s -S tt Upig tis- "II at tsll nmil an a ta l a m bI' e1 eN it n wt tts S'S * S-l m-l#t' S$3 il-1 - 'ttl st t-|-e retl'l 11 trt am a It on ui tt It w U UV I of aI lt nsre CII S' Isla et ,:Ii uci : uu:Is s) Ut of's A It an a IX I 1S0 I am fsat * mt , nstt s' ull eI's It't #t 1-eI I,10- SU t * 11 calm an |~~~~~ "Is a a s a lom tii ul aI§0 Is at N V I,1' 1 11,1 Is-et," $t'I$ e- s Mau-S t'l' ttt't t'l fall Hll "Int 1- o t t aus 4 f wit Iti i It in ii K Im1'a 1#'1 rit loon a': L-a aa1 M.t WIN I,- utl as'u Wil e-n a" 3335 n 13 1 5"'l . ft " 1 rtll40 t1 5"1 t *- 0" i a aa esi a Isn'sal,u Is II . usu'm I'S IIs as;1 titlems tgom ul hits U-n womi m a a N W U1*0 I at a is ii t' s tII $,I : cu'll I:l a-s snt Iuol tsa'l s-i uI' fl mor am sn_ n t az na a tt ' r i I# neul1l oil In z1 'll AllU wl- ttt M' fell's 411 m o np a ms us * s u, at nII se IUgnI !1 9g In ma 1-1, $ I- t g I We * aun a t gMI itt A sa 1u,'s re II it sell I I It u Slln's 051. sl WHlot W S " ofU tot au it '. t-l re la liU 65 ii 3-t ihl tl'l 1t ult Ut t36 MIm till"'tt Im' tit 1-1 t-l t stl't It lit tal tus'a M'U 1 h: s ta tiMi IUN rn' ItsuliII tsa i tm Os'; s st- ii6 ma's; Lst Iis as muit lull) I,u- fullt aM int nu= II= U It It at Ita M It it s U i ta li Is so t-s M'I: IS III nue t l'i hUil sMt) M t'l m "aM ni It oil A it of M 4 m A 0l 5-t ittai'ttueI rs l i9 -ii 33 U'A 1 't 3M 51 11it VI= U # U tlUU'I a: nt, *In tu iet A- 3 U'i ilt aa-tt tas'sU .5 n' 5111 sts I)is s'm stS - n t ~~Mst onlN It's I '1eU'n sa nr rl ntXl" II I'S 13 Was "I'm III "I'll let itU'I#"- Ml ant ItofIt "1M It 3 It1 UIt I It A t't I-l ii rot "ut r st a r nl r sa mWst ii fam al sl M nall am mot "I n #3'll tl r rs 1.e1s3 Is i l t I I UttU ml," . s in a te i't _ _ M a a A it a 'Swt re 55 S-i AaA rI au flo $Wu Si't II'n e. ts t U-t lM _m a la a amn tol I Uo U'I wil ae It nztmlt II n. vt flt 91 II I' to" on r 1 ala IUl u . U n6n63663 t Ut~ nW Uli A ttt't it- 1 i -tt O'S r.3 t-HO U "§"t *n 3U'5 35 05@ _ii i_ ast n Iss a tM A Olt ita w nra II I'llt In's $1 I't a' atl'l Wi *U't t Nsti frll _ W 101 W 1" {t) 11*1 I' re ro Wo 0, A' ft z ^ m l 1 1 ml1 m-"1 #1 on1 "I -I(l I(s mlWNSW (lWII WI -J) I'sart fll i'U s ml etF * * m *l-tl won tnsail *lo a "l *-'5U-- n _I -11/h -5313- i 'a 1 Iain -,11_ .- . ma ---mu.----- own.; m u uN o a1 am aWlin mum- at mmiluwm nun.. wAU= wanm luEMi- ----UmU 61d Ill -- ---mim amii au---- -mum U ----mm "a--- as s wF- (Mum) (tZecmisumoa m au mmo U _ tw :M M tsm 4; oll ai as I4i 5 3613 n I1 SI *6 635 II WI Usl 9 AVI'n, SWill s'stt V I-l rttu uWt*ut nlriu G l sl u | | e I el Ins 16n 1 iu ussl' ri gtt em. £F 61 s-n wsu *Ws rt ut1u wt ismim urn wt n s m,i sims': a ant n et ttm rs r u mn VI :i t t twn otwl rt us II-t s a Us n *ISU Pila n N' n Eus a III a a P Wet 33r re *1 e'IS I' It i's SLI£ Nl53 a6 t III 11f s oi 6 5)PU WI Si 9 n ow ,nt WS N 6 11 3nhs'St ums urn Us us aS I ls' Iw i w1:' s n ita swtr re 1- wo a r s tZews ars *1 0 |W 'sl alail 1l umtu Hln- ataI a esse usls rt re SUt 603 vt rs rt ussas's s I i s 53 V a a a 6M's I a Uli L a 5111's 50 £5 IV IS% 1n 1 30 11 L1W' n'l twit £ i "I WIS f n it M si ra a a o e 's rs 53t t rt ta t"l Mt t tf m finu u I m"l a a a twit In m I 0 n Is um" s rl *-U in, O3 t V 1is MINI wt# m1 mt tm a U u1 3t 5511t 61 U' 5U n6, U n3 56 66 51 56 11 tss R51 L1 3s U t*N UN 6t WIt 51 WP *fl1U 4 I1 us at m z$nlt itt tS ISt off a S It as VIt III re g &3 I' u 14 l":to t "' us "Itt"$ amfa __ ~~~~~mts nt'n re B 1,1 rIlt'l_- tuou' III * r: 91 s wi IIIl Uri Bw *tl vi ISA) UUU S6SU SW.". 1151 !s'N S"tI riot. sus *, ** 61 SWIl ar_ {t@ Ut _I 6 IMIIUM i_ a aNKs'S 116 1111 u a K mn I tar us £1 I I's II Wm lull I1 a _ 6,, M11 a a i aa onofWsa LU' I I 0111 I's 56 I't, 011`5333 St i V11 Swin MI1I I9 W' full 6631 13113 a5 atU II a3 4111:41 us' ow K' 0,14 N et *ul I's I's r3 MOM' gI Ii gI's 1111'Il Llt of Ut InUs sla WmU iu It) I) IS)WI Il 1316is) 1o) II) IN) I)IS) I)15 UII 131oml=l m ami ml am (a) list 3161 IU66) I --WU-- 4 6 S Isefil -6113- -415/- 'A . '3 a3m -Ed41 II- 15---Ed U-WWI 51l WIt 0135 itS Ml5 1U 55a .wiaN liust us whmmmi 011i11m1 Ulm. lAU= ---3 3 s1- 1---111191 fluG - - U 3 ---- -----MM - Ii.5USI 15(ll) aSi xu11fh U 1WU I :g UW t.11211U1 0 t1j1490A "D *OUVI 4*90 P"wl to to t * *'_s i *-- l ae..s3w as so 2 s ae (4 *1 Xa.§sssS - SUd M3tp* @1 Sq3asgao a. *- 1 Sb Iw ---I- -- X------ ------------ 6* 'US 1 ----9---U- -------- -- - uses 1 8 1 --a£--- - -.. . .. . t ^ ** _____tt____------------ V go toa m 4f3 9 *- ----SI----- * - -----9----------- . Vs out *- -- ---9--- ------ fit 4*13 ___0 ----_-- I 4-.. Me n 3 dl 15 '. 091 *- t ---di---- *--_ -is-t__-_- --__ _- ---i--- 11 *a33@ to .95 9£ W1 9 ---S-t-- *- --------- - *- 9 * It 11u93 5093l £ --11--tF ---------- *s 95 i, * *a of -t on ou *- __j ... S-t *| 10411 9E 9 Lu WI * i- -- 9-- ----t9------s----- is *s3* sf* W$ I I ----s -- -------- ---- --- -- 9---to----' --0t 3 9* @3 WI 013t -- --- £ ---* *------- to- 9t _t of SI tt Wo t It U s - 9 1 t ----1£- 1n SWS 33 13 II I WI £6 £3 ----19---_ - -IS---'-------- '--- --it--- 111 *3P, us IwsS 1 S 39- ---_ 3-98 U_ .* U t391 east -----as--* ---i---t III is wasmi0I £ .3 331 *~, 15 ** it I -- 1--- t ao iU WS1 9* 4,-i w.g S * i ----iS--- _s 5_ . 3 iSt *a"t 9£ * lUst 0 *- ----1----* ----3-- *- *- S- #-- - 93 93 ut wI *- *- n * --*-g ------t- dl--- 6* _ *9 95' 191 63 Si *- -- , ---- *9i- - * - ,,,0 "u -V {fl /ItS 159 OI' OS s ---IS----* --- tw--- *- *- ttw oil it to to@ _ -___1_| ,w *1p9 4 WIz w 3£ ---i- ---a-- at lo *ap-ots -.u * . -1 *9 ------- ---s-* * 63to is DII to _ ;I 1- - _ 01 S 603s136 98 g *X 9- WIU* *---4 - ---9 -----3 It" £e1 Wit 1563s -. ---1 -- 3 $ ----a i3 1 3 * *~~~~~armps " so W _ W q, ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ U salzonoa .aue.eg p} sgauaw zwasooo oi, us 3606I s*J2u@DPS3ODIS( UT SU.nuAeu 3U*iUOAOD 03 UOT31q5;%U03 pow am3n.8 sqU :BszJ, jams pVon :5 *qqj 103I181232S J I Amexn -~~~~~~ uanni uni ain * ~ ~ ~ ~ ~ ~ ~~o u W - Ilum Cot-f" fetveO_im for oa Zawatmntand Moaitensane IL TheAma,teal Irmaak 1. The anlytical framwrk for the teechical flndings Is the World lank study on highway desigp and maintenance stanards 16, 27, 291 and the oprational Highway Deig and Nalatenance Model (lUK) developed from It (now In Its third version, K-III). The model enables Investigators to apply the results of empirial research on underlying physical and economic relation- ships to evaluate pollees, standards, and programs of road construction and m*atenace. 1421 The geoeral relationships aMong the main variables ia the road deterioratlon cycle are shown schematically In lax Al. The model contais statiletically validated empirical relationsbips that are used to simulate aost of these effects. Using data on existing roads and their conditions and estimaCas of future traffic volume od composition, the model preicts the deterioration of roads and the effects of specified road mainte- ane an lprove t policiese modelG then estimtes the cumulative effects of the" processe on road condition, particuarly road roughness (one of sevral road condition parameters estimted by the model), which has a major effect on the operatlg costs of veheles using the roads. 2. Tho model prdmts the evaluation of several maintenance alterna- tives for a road 1lnk, a group of roads with similar characteristics, or an entlre network of pved and unpaved roads. The moel computes the aggregate costs of carrying out specified maintenance and construction polleles, the asociated vebtcle operating costs, and the tin streams of total life-cycle costs discounted at several rates to find not present values or internal rates of return Using these decision criteria, optimal combinations of road design and maltenane policies are acortained. MWanteanc policies under budget constraints are foud by linking the EK-Ill Model to Its companion Expeounture Budgpting Mdl (W), which employs dynamic programming to deal with multiperlod resource constraints. 3. Vehicle operating coats constitute a larg share (75 percent to 95 percent) of the total road tramport cost, except whAm the traffic volume Is extromely law. Thus the effect on total transport cost of even a small percentage change in vehicle operating costs Islrge relative to the effect of changes it construction a"d maintenance costs. Before the MU studles, / Thbis annex prents a fuller techalna eWplsaation of the findings reporte i the mai text on road lnvestumnt aNd maintenance options and tbeir economic cequences. It is based on an lanvestigation of the englnerig-econom c aspects of road deterloration by handari et al. 141. - 56 - UU?UNINAMTS 70 Of8 0? QtRiORATION - topography and subgrade oil - material *n4 thieknes of aved roads unP vd ro pavement layesn - potholes - potholes - dralnage (surface and Internal) control the - ruts - ruts - quality of constuction progrsooion - cracks - corrugattons - environment (rain, frost, solar of - broken edges $ evosion radiation) - shoulder wear - gravel loss - traffic (volumo, axle loas and > dn e t form.- configuration) tion - existing road condition replace ensur worn-out II adequate surface draiag ,,result . . . _ ~~~~~~~~~of F N ZN WI C. wate . - keeping drainage clear - patching CraCke ad potholes C aOUOESB - repairing shoulders With -J effect on - ressling/surface trestmeat - pavement strengthening/ rehabilitation V8HICLE OPUATtNO COST - reconstuction (deterred s pard - ep parts saintenance) - fuel consumption - maintenance labor A1_ - tire .sr vehicle utiliza- tion and age Unsaved ro"Is: tn the ab4sece of maintenance, unpaved roads detertorate very rapidly, reulting In vaty bigh levels anm rapid growth of roughnes. The progression of roughnessf,howver, is eeutially linear, aNd th consequences of poor maintenance bocome apparent imediatelr. TS action of traffic causes corrugatonas aod rutttig and, combined with rainfall, results In deeper ruts, potholes, and even gullies and washouts. -xcessive water reduces loa4-bearig capacity, itduciag rapid deterioration under traffic. tn very dry 'onvAtions, evaporation of water reduces the bwod witbin the surface material, which then diointegrate under traffic, so that the fine binding particles are lot io dust, leaving loos gravel, which need priodic replenoshmt. The wear and tear on roads by traffic io thu tronogly lterrelated witbh clietic factors. Topography an and roAd alignment ar sigifleant factors because eroso, vehicle cornering, and hill climbing tend to increa gravel displacoemot. Se conditon of gravel aod earth roads is hSghly sonsitive to the ldvel ao ftequeny of Maintenaca. Paved roads: Th progretsio of deterioration on paved roads follows a distinctly nninsar path. A long lntial phas, tlsting up to two-thirds of the life-cycle, passes with little discrnble distres and minimal maintenace osn s but to followed by a phase of accelerating deterioration (primrily In the form of cracking and rtting) resulting In itnceased roughness ad, o extreme cases, potholing. tn the absce of a major maIntenance intervetion thi eventualy leds to structural fallure* Whil crecking coaus a generallsed rtduction in pavent strngth this effect s exacerbated by lnadequte drainage ad shoulder maintenace, a rcai water penetrates the eracks and wakens the undetlytig layers. Pavent d ge due to poor daing Ls particularly pronouncd in wat climtes oa In aras subject to f aing. tn ono ase studied here, heavy rainfall (l5O0/month) reduced pavement life by about 30 percent (ile., frte 14 years to 10) compared with light rainfall (20 mm/month). pavement deterioration is critically lnflueced by traffic level an loding, with d Jma Increasing exponnatially with ale load; the damge, however, may be curtUild by tacrasing paveent strength In term of its structural number (). e d oRn typicl eSxmples LA this study (pavemens with en initial SW of 1.5 to 2.0, ed 0.25 to 0.45 million standardale loads a your), doubling the initial strength pot_pone the mad for an overlay by 10 to 12 years. - 57 - no basic empirical data on the relationship between vehicle operating costs and pavement surface condition were available, Road Investment decisions were guided by Imprecise and fragmentary estleates of vehicle operation costs related primarily to the type of road (paved, gravel, earth). Based on Improved estimation of vehicle speeds (In a free-flowing traffic reglme) and operating costs, as a function of road design characteristics, the lDM research suggests that vehicle operating costs are somewhat less sensitive to changes in road condition than previously estimated. A comparison of vehicle operating cost Indices for paved and gravel roads sbows significant differ- ences between HDM results and estimates from two earlier standard works (Table Al). 4. The relatively lower vehicle operating cost differences estimated from RDN relationships may be attributable in part to Improvements In vehicle design over the last 20 years. Nevertheless It Is faitly clear that certain road improvements may have smaller benefits (vehicle operating costs savings) than previously estimated. Inadequacies In empirical Information on basic road deterioration and vehicle operating cost relationships In the past may have contributed to premature paving of toads, particularly in the low- traffic, high-cost African enviroment, and possibly to the underdesign of heavily-trafficked national roads In India, Indonesia, Nigeria, and Pakistan. Tabe Al ComParison of Vehicle Oewratin Cost Indices for Paved and Gravel Roads by Vehicle Class Passenger Car LUght Truck Reavy Truck Articulated Truck Paved Gravel Paved Gravel Paved Gravel Paved Gravel de Wellle (1966):/ 100 126-133 100 141-152 100 140-170 100 150-170 AA8HTO (1977)b/ 100 108-163 100 109-156 100 112-158 100 113-161 nDi (1983)0/ 100 108-126 100 114-138 100 117-146 100 115-144 8/ Based on de Welle - Quantification of load User Savings, Table 11, pp 28-29 1101. b, Based on Winfre - Economic Analysis of Highways, Table A-44, pp 727 1431e C/ Based on UDK relationships applied to Costa Rlca data, with steady-state roughness on gravel road,ranging from an IRI of 5 to 9. - 58 - Case Studies 5. The 1DM-111 and DM models were employed in a series of case studies to investigate optimal investment and maintenance policies under a variety of conditions. load types and conditions, traffic, climate, and unit costs in the studies corresponded to those observed In three countries Mali, Chile, and Costa Pica. Despite the limited inferences possible from the case studies, the consistency of the results permits generalizations for wider application. 6. Thirty-one maintenance policy alternatives were tested for paved roads and ten for unpaved roads. The policies consisted of different mainte- nance packages with specifications of the deterioration levels at which they would be applied. The pavement maintenance options ranged from low-cost pothole patching, through bituminous resealing of the entire surface and more costly asphalt concrete overlaying, to major rehabilitation and reconstruc- tion of the base and the surface. (See Box A2 for a classification of road improvement and maintenance works.) As a bonchmark against which to measure differences in the costs of other alternatives, a 'null case" was defined that included only the basic routine aiatenance activities (drainage clear- ugs, miniml vegetation control, and shoulder repair) that are also included In the other alternatives. Table A2 sumiarisea the alternatives, Identifying each by a code (such as ALZS), which will be used later in discussing the results. To illustrate, under alternative AL18, 100 percent of potboles would be patched each year, a surface treatment would be applied whenever 25 percent of the area was visibly damaged, and a 40-millimeter (am) overlay would be applied whenever roughness reached a level of 5 on the Tnternational Road Roughness Index (I U) scale. 7. The roads in each country's network were grouped Into broadly homo- geneous classes according to surface type, condition in the initial year of the study, and traffic volume. -For each class, road deterioration and main- tenance activity over a 30-year period were siamlated under various mainte- nance alternatives. Road maintenance costs and vehicle operating costs were computed, discountei to the Initial year, and subtracted from costs for the null case, giving the net present value for each alternative relative to the null. From these results, the best strategy for differeat levels of avail- able funds and discount rates was determined for each road class. The results for different road classes were combined with the aid of the Expendi- ture Budgeting Model to find optimal network strategies, costs, and benefits under conditions of no budget constraint and also under varying levels and time periods of overall budget constraints. The detailed results from the cae studies were used to explore cost-effective road Investment and mainte- nance options. The main findings relate to the following factors: * Selection of cost-effective maintenance policies. * optimization of maintenance expenditures under budget constraints. * Decision criteria for pavement strength. * Economic traffic thresholds for paving gravel roads. * Impact of overloading. - 59 - A alasfeLaUs of _Bt ntmoe t Jumveinm w6bho Routine Kaintenance. Locallzed repair of roadway and pavement,-gradIng of unpoved surfaces and shUlTes, rogular maintenance of road drainage, side slopes, verges, traffic control devices, aot furtnture; roadside cleaningt vegetation ad dust control, snoc or sad removal, and Maintaining rest areas and safety appurtenances. Typical coats range from less than $300 to over $5,000 per km. Resurfacing. lesesaling a paved road or regraveling an unpaved road to preserve its structural Integrity and ride quality. A paved road normally neds resurfac- Ing at the transition from "good" to "faitr" a=dltion, provided the volume of traffl justifies retaining It In good conditlou. lesurfacing is sometims called "periodic maiatenance," even though all maintenance activities are periodic. Costs can vary from under $8,000 to over $40,000 per km. Rehabilitation. Selective repair, strengthealng, and shape correction of pave- omet or roadway (includinSg minor drainage iprovments) to restore structural strength and ride quality. The term stEroheonit is sooetlies used to describe a specific category of pavement rehabll1tatioi iiiolving the application of over- lays. Costs of rehabilitation can vary from under $30,000 per km for unpaved roads to over $200,000 per km for paved roads. Th costs for paved roadso, how- ever, rise steeply as a pavemet deteriorates from fair to poor condition. Reconstruction. lnewal of road structure using for the most part existing earthworks and road aligment to remedy the cosequences of prolonged neglect of malatenance or where rehabllitatloa is no lopger technically possible. Costs are quite variable, ranglng from about $45,000 to over $300,000 per km. Restoration. A generic activity encompassing major rehabilitation and reconstruction works. Betterment. Road Improvements related to the width, allgnment, curvature, or gradient of road (Including associated rehabilitatlon or resurfacing works) to enhance traffic capacity, speed, or safety. Bettermet works are not considered mainteance activities except for acillary road rehabilitation or resurfaclng operations. Costs are bighly variable depending on the geomotric Improvements. New constructlon. Constructlon of a pved or an enginered gSavel or earth road on a nw alignmeat; upgrading of a gravel or earth road to paved standird; provision of additional lane capacity; or construction of sa additional carriage- way, frontage roads, grade-separated lntercbhages, or a multilae dlilded high- way. Costs of new construction can vary from uder $50,000 per km for a gravel road to over $1,000,000 per km for a four-lane access-controlled divided highway. 9 ~ . . . ,. , ., , , , ,, , , , - 60 - 23SU3 A NmluLe ?ollcv lemugel"S t L avedi d a d 1a / A. PAVED ROAS PATCHN ALTnUNTiv (2 ARA OF mMUa PO)OL) uszu.u ov0A c ALOO (Nall) 0 AW 50 ALO2 100 AL03 0 330 if AT 6.5 31 A04 50 1C 8S Al 8.5 Ut ALOS 100 uc ST? 0*2.5 3 *1.0o 100 ST 82 lsx Area u sdb AL01 100 ST AT 50 AresOD £1O6 100 S A 25t Area Daged AL09 100 ST AT 75X Are Dued SSme i A6.5 3S 3 A10 100 ST AT S0 At" hmaed 3C a Al 8.5 3U ALII 100 ST A 2532 Ae&mge 3B0 Of A? 3 8.51 AL12 100' 40 _ A? 3 SU *L13 100 40 n t ?4.2 tt AL14 100 40 = AT 3.5 IU1 AU.15 3W OVEA (so0 u) + AL 12 AUG 36 OEA (80 m) + ALT 13 AL? lI oVLy (0 m) + AL? 14 A$18 100 a * 25 Ae Damaged 40 a At 5 tlk AL*19 100 ST ATS 0S Ana DMg 40 AT 3.5 x £1.20 ttO6 IUC 8T. *1.12 *1.O 316 330 ST * ALl) AL21 Slll inI ST + AL13 A122 TM Wl1U STf *1.14 123 0 3U AC A? 8.5 tU A24 0so 0.3 AC A? 8.53 3 A£25 t100 M3 AC A? 6 £126 100 ST A 75SZAC" Damge me AC A? 8t.53 3 * L27 100 ST SOXAT" bmuS.d 33 AG A? 8.5 SU AL28 10oo ST A252£teamaged Mic AC A? 68.5 AL29 36 AC o A13 ' AL30 3 AC* £11 £A.31 KU 33 AC + *11 ALTERNATVE ULADIN IUzqIJNC 510 22IOSUVL GRAVEL 335U3AC33 RUN= £1.00 (Nul1) bms NbeW AWI own a you 302 of terial Loa NM A,02 vey 6000 gm . ?aee" 'bs ALW3 iv.y 6000 f'e Pms" Wu A£04 Evaty 4000 Veh fte" you AL5 Evey 2000 Y. . e N"o AtO -L10 Se as A3 -AIM but with Grael AWbts cig 150 mm ukesteve thdcms < 50 a / All alterntives tInclu e otlm uiatene daia aring4, ants" vegtaties coutrol, repdr of eholdar VW drin. bgamae area coustitus the area of pveut srface with speciie *ig of diateem md Visible df ectOs Teraologit SS a bitamioe surfoce treatmat AC aspat cesm MNC a rconstructiom with lither ST or AC surface t16 a lm iate (looe, first yea of the anlysie period) 2 . Itemstionsl ouas ms atde - 61 - Selectiom of Cost-ffective Imiatm Policis 8* The choice and stagling of maintenance operations on paved and unpaved roads are strongly affected by the differences In their deterioration characteristics. On unpaved roads, the linear but rapid path of deteriora- tion requires special attention to routine maintenance, particularly the frequency of blading. The nonlinear deterioratioX characteristics of paved roads offer more options for the choice and timing of maintenance. 9. To Identify cost-effective road maintenance policies for paved roads, particularly under budget constraints, the trade-off between agency and user costs was examined. To Illustrate, Figure Al shows the results of such an analysis, covering 21 maintenance alternatives for low-volume paved roads in good to fair condition In Mall (average daily traffic of 400 vehicles),.i/ Maintenance alternatives and their specifications are Identi- fied In tiTs and other graphs by the codes In Table A. The bar cbart (Pigure Al-A) shows the present value, discounted at 12 percent, of the net cost savings relative to the null ease. To arrive at not cost savings, the increase in maintenance and construction costs incurred by the road agency. 1 subtracted from the difference In vehicle operating cost. In most cases this yields a positive saving, but alternatives L.15, AL16, and ALI7 (Lavolving immediate application of a thick overlay) and alternatives AL20, AL21, and £L22 (involving immediate reconstruction) cost more than the benefits they yield when dlscounted at 12 percent. 10. Net present value, discounted at 12 percent, is maximized by strategy AL12 (patching all potholes annually and applylig a 40-m overlay whenever pavement roughness exceeds 5 IRI). But several other alternatives with very different combluations of vehicle opeorating costs and road agency expenditures, are almost as good by this measure. This finding Is important because It widens tbe room for maneuver when agencies are subject to budget constraints. For example, the curve In Figure Al-B relating agency expendi- tures to vehicle operatlg costs for selected alternatives shows that AL05 (pothole patching until roughness reaches a IRI of 8.5, than reconstruction) entails about half the road agency expenditure of alternative AL12, while vehicle operating costs are higher by about $1.10 for each agency dollar saved - a net loss of only $.10 per $1.00 of reduction In agency expendi- ture. If funds are lilmted, this may be an attractive option. 11. The net presenk value of total transport costs for different levels of agency expeaditures is illustrated in Figure A2 for four cases. The most efficient maintenance alternatives lie on the positively sloped segment of the outer boundary - the "efficiency frontler." Alternatives represented by points Inside the frontier are always Inferior to a 2/ Results proved to be Identical for some pairs of the 31 alternatlves; for such cases only one of the pair Is graphede - 62 - X Ret P?mest Value of Maintenance Alternativs 40 - 30 - f ,, ._. ____ 20 U I Pl)-~ ViSeeXi 4bSw i4.. ,S 110 -30 -50 0U o I to4 i Ia 'I. 'ii I (B) Vehicle 004ra.tinx Costs Vensue Road Agmey Cost. £10(uALl alt.) AL.02 (ptebtas only) Mm.0 (patcbLn 6 nsrecaSucttmi) JA*1.2 (Oetlam St S tRI) US0 *1.3 (overly at 4.2 11) O ~ ~ ~ vu ses= Zwad ('4l0' u# per km) Figure Al: Low Volum Pavd Road in Good-to-Fair Condition O:ali): Analysis of Maintenace Alternatives 3 63 e 4h 1180: 2C~~~~00; ' t0 ^.~~~~A ..... OAL00 *ehlel a by " . llouo 703hllsB *4 403~~~~~~~~~5 C.~ ~ ~ ~ ~ ~~~~~~~C 0 4 g *3 14 20 24 25 0 40 00 so ftt"fAvcc Rmeav.OAg19Mv Cam A. Low-volum (400 vehicles a day) paved S. iedius-volum (700 vehicles a day) to pod/fair condition (Nell) paved rad In poor condition (CHILE) (Cout~~~A 14" (Coot Uc ) 3000~~~~~~~~~~~~~~~~8 350 30 !2oo~~~~~~~~~~ p300~~~~~~to so.~ ~ ~ ~ ~ ~ ~~~~~a 10 a 10 20 30 40 50 to 20 '310 40 50 5 PeMMVES.n 90co Coe bggeWi a AgefIC Cam C. Nediump-volume (800 vehicles a day) D. 34kh-volume (1500 vehicles a day) paved road In fair condition paved, road in poor condition (Costs Rica) (Cost& Rica) Figure A2U Nt Prset Value (VW) Verss Agecy Ependlture for iffeent MIlatenace Strategies -64- combination of maintenance options lying on the.frontier. For example, In Figure A2-A (Mali), alternative AL09 has a net present value of $12,000 a kilometer at a present value of agency cost of $15,800 a kilometer. A higher net present value should be obtained at the same agency cost by using AL05 on part of the road group and AL12 on the rest, giving an average result that lies on the line connecting ALOS and AL12 on the graph. In the absence of budget constraint, the alternative that maximizes net present value should be selected. Agency expenditures beyond this point (AL05) result in benefits that are less than the Increased costs, as reflected by declining net present value. The curve also shows the optimal order of retracting when agency resources are cut. The lower-order options arc selected by successively reducing agency expendlkures in ascending order of their marginal contribu- tion to net present value. Where traffic is heavier and especially where roads are Initially In poor condition (Figure A2-8 and A2-D), the economic loss per dollar saved by the agency is considerably higher and, of course, successive reductions In expenditure have Increasingly costly consequences. The steepest slope of the efficiency froutier, lndicating the severest penalty, Is reached wben the only expenditure available for cutting (always excepting minimal routine maintenance) is pavement patching. In none of tbase choices is agency cost today traded off against significantly higher expenditures in the future; costs are simply traasferred fon the agetcy to the road users. 12. A sumary of recommended maintenance options for paved roads, as generalized from the case studies, Is presented In Box A3. These recoomenda- tions reflect interactions between traffic levels and road conditions, and the aintenance Interventions needed to reduce the total road transport cost. In indivldual cases, the level and frequency of optimal maintenance actions may vary significantly from these recomeandations because of varia- tions in factor costs /, road construction maintenance practices, traffic loading characterlstlfs, or envlronmental conditions. 13. On unpaved roads the primary maintenance-related determinant of roughness - and so of the cost of operating vehleles - is the frequency of surface blading. Simlation and costing of the effects of the various main- tenance options (Table A2-J) show that blading costs ad vehicle operating costs are closely balanced over a very wide range of blading frequencies. Optimal blading frequency is In the range of one blading every 4,000 to 8,000 vehicle pases which is consistent with generally accepted good practice (for example, one bladlng a month at 150-200 vehicles a day). Incremental increases In blading frequency within tbhs range result In marginal reduc- tions In user costs that are almost completely offset by corresponding increases In agency cost. ligure A3 sbowe plots of not present value versus blading frequency obtained from the Costa Rica case study. Iven at low 31 For example, the results of the case studies suggest that surface treatments becom a viable economic alternative to asphalt overlays when the cost of surface treatment drops to 25 percent of the overlay cost or less, on a unit kilometer basis* - 65 - * 01110W~~!M 0 IIZ TOTAL Xos~~~~~~~,g 30!m we "ZO COST0O PAT= &OM3 Average Daily lnitial Traffic (ADT) load Recommeanded Maintenance (vehicles) Coadition/ <200 Good to Pothole patchlig only until roughness level Poor becomes very.high (8.5 IRI) and the traffic volume has risen sufficiently (ADT > 200) to warrant pavement reconstruction. 200-500 Good to Pothole patching and overlays e.g. 40-m asphalt Pair concrete (AC) whenever roughness reaches 4.2-5 .II. 8urface treataets may be substituted for overlays under budget constraints. Poor Imediate reconstruction of the pavement-b 500-1000 Good Pothole patching plus periodic overlays (40-mm AC at 3.5-4.2 IU). Fair Same as above but with an initially thicker overlay (80-m AC) in cases of weak pavement. Poor IMmediate reconstruction of the pavement.b/ 1000-2000 Good to Pothole patebing plus periodic overlays (40-mm AC Pair at 3*5 Il). Where existing pavement is week, the initial overlay should be thicker (80-mm). (Periodic surface dressings are economical on strong pavements in good condition). Poor Immediate recountruction of the pavementb/ >2000 Good Pothole patching plus periodic surface dressings (at 25Z damaged area) in addition to periodic overlays (40-mm at 3.5 IM). Fair Pothole patching plus periodic overlays (40-mm at 3.5 IRI). Where existing pavement Is weak, the inltial overlay should be thicker (80-mm). Poor Immediate reconstruction of pavement.b/. V/ Good - II less than 3.5 - Pair - IU 3.5-5.8 Poor * IRl greter than 5.8 so.e provision for patchnlg and emergency maintenance should be made to keep the road serviceable during the reconstructlon period. - 66 - 5OO'. _ J , - ---_ | 200 AOT s oO '00 ^'~~~~~ ~~~~ ~ -,a a V=o os100 1.0 1a.m 41 VehiLcJ. oa". between bLadings -i .o : - - blue hI~~~~~~~~~~~~AO *al u11S2: 7igiw £: 'eiaL. o Nt teeu Vlu __t- ---d--g frequ==c - 67 - traffic levels (25 vehiles a day), the economic returns on blading are substantial. Ilading once a year appears to be an acceptable minimum threshold. While blading once after every 4,000 to 8,000 vehicle passes would be an optimal polley, less frequent blading does not occasion serious economic los1, If the road Is regraveled at appropriate Intervals. Local conditions will strongly influence the maintenance options for unpaved roads. Extreme combinations of soils and climate (sands In arid climates, heavy clays In wet climates) may justify surfacing earth roads with gravel at an average daily traffic flow of fewer than 50 vehicles to ensure accessi- bility. 14. Where capital - especially forelgn exchange - is scarce and unskilled labor Is available at competitive rates (currently about $5 or less a day), the use of labor-based work methods for spot repairs on gravel roads and of simple drags attached to agrlcultural tractors or trucks to even out surface corrugations could prolong the time betwee blUadigs with mechanical graders. For unpaved roads with very low traffic volumes, increased use of labor for spot repalrs could offer a viable alternatlve to mechanical aeanp of maintenancs, with the lower maintenance costs offsetting the somewhat higher vehicle operating costse.81. 0etLiistion of Maitenance lzsemditmres boer hisdtE Cematralts 16. When the maintenance budget for a aetwork is less than that required for the overall optimum, allocations to specific road classes4/ must be redued from their optimal levels. towever, maintenance expeattures should not be reduced uniformly across all road classes. The loss In net present value will be ainmlxied by first reducing allocations to those road classes for which the efficiency frontier (para. 11) is least steep. In general, candidates for cutting are high unit cost malntenance operatlons on roads with low volumes and good existing surfaces. After the optimal activi- ties have been completely replaced by next best options In a road class, the penalty for further cutbacks will be proportionately greater - the frontier will be steeper -ad it may be economical to reduce allocations to some other road classes as well. Roads with very higb volumes and poor surface conditions suffer the greatest loss in benefit for each dollar of reduction in maintenance outlay, and their allocations should be reduced last. 17. The effects of reductions in the sainteosance budget on the choice of alternatives and on the resulting benfits were exalmied for the road net- works of Costa Rica, Chile, and Mali. In Costa Rica the maximiation of total not beonfits would require spending an average of $12.5 million per year on maintenance over the first 10 years with $38 million needed In the first year alone to rehabilitate the paved roads In poor condition. Figure A4 shows the maxium nt present value of benfits (discounted at 12 percent 4/ Lengths of road sections grouped togetber on basi of similar 4ssign, traffic, and condition characteristics and, where appropriates, further stratified by reglonal, climatic or other relevant factors. - 68 - eo0 Costa Rica 41 _~~~~~~~~~~~~~~ : I t~~~~~~~~~~~~~~~~~~ 200U /aI 0 , , - , I. - -. ' !- -. ---- * ~ ~ ~ ~ ~ sI~~A o AewiM MowmflsIw keAMMn (10 Yr- Av% -~~~~ mm-a F71m Ms Vaulation of Not Preet Va<" for tXe hdtirs lad Netwo*t an a Tuetia of Average Annul lainte_ane hpendtre - 69 - a year) for the Costa Rican road network under different levels of average annual maintenance expenditures, optimally allocated among different classes of road. If the budget is raised from the presently planned level of $6 million a year to the optimal level of $12.5 million a year over 25 years (or by $51 million In present value) the attainable net present value (with the best use of the funds in both cases) Increases by $200 million, from $635 million to $835 million. The optimal program, even with unlimited funds, would not keep all roads La good condition. In Costa Rica, it provides for a high level of maintenance to keep two-thirds of the paved roads in good condition with the other third to be maintained to lower standards. 18. For Chile the best maintenance strategy for the road network is silmlar to that for Costa Rica: two-thirds of paved roads to be kept in good condition while maintaining the rest at lower standards. The optimal program Involves immediate maJor expenditures on the rebabilitation of paved roads and assigns high priority to that activity even If the budget were cut back. That is the course Chile actually adopted, aod the planned budget for mainte- nance of all roads - $140 million a year during 1984-91 - is approximately the amount required for the optimum derived In this study. If this budget were reduced for any reason, the losses would be large for tbis highly trafficked network. 19. In Mall, by contrast, where 84 percent of the paved roads carry less than 200 vehicles a day, only patc'ning and basic routine maintenance is economically justified on most roads. The most economical option Is indeed to keep only about 1 percent of the paved network in good condition with the rest of the network maintained at considerably reduced standards, mainly by patching and routine mintenance oan paved roads and minimal blading on unpaved roads. ven so8, about $9 million is required to clear Mali's backlog of economically warranted rehabilitatlon projects for higher volume paved roads currently In poor condition, And the average annual expenditure required to maintain the combined network of paved and unpaved roads is esti- mated to be about $6.2 million a year - about twice the current expeadi- ture. If expenditures were raised to $6.2 million ($24.3 mdlllon In present value over 25 years), the present value of the benofits is estimated at more than $46 millon. 20. Maintenance activities are often deferred during periods of auster- lty. Tor unpaved roads, as long as basic routine maintenance is carried out regularly, the prlmary effsct of deferring blading and, to a lesser extent, regraveling Is to Increase vebiale operating costs during the deferral period. The effect on subequt road restoration costs is not large unless the road Is allowed to become virtually lmpassable, so that It has to be reconstructed, generally on a new alignment. for paved roads, on the other hand, both effects can be Important: vehicle operating costs Increase during the deferral period and- the cost of later pavem8ent rehabilitation can be increased substantlally, depending on the stage In the deterioration process when deferral occurs. On newly constructed or rehabilitated pavements with light traffic loading, the effect of deferring malntenance (other than basic activites such as drainage) for one to flve year is negligible. Once pave- ment condition becomes fair or poor, the Impact is large. - 70 - dlIsi Criteria few Pavement Strength 21. When a new pavement to constructed or an existing one replaced or wverlayed, the choice of design strength should take account of the rell- ability of future maintenance. Low probabilittie of adequate mainteasce and timely strengthening In the future favor building a etrong pavement Initial- ly, since stronger paven 4ts enjoy a longer grace period during which mainte- nance needs are minimal. / High probabillties of good maintenance will favor time-staging - that isg,ecofizag on today's pavement and strength- ening subsequently as needs eerge./ 22. To justify the time-staging of road construction a minimum prob- ability of adequate future maintenance is requlired. To estimate this threshold probability, life-cycle costs for pavements with Initial structural strengths, (SN), of 2.0, 3.5, and 5.0 were estimated for a range of traffic volumes and axle loadings in Costa Rica (light) and Mall (heavy). Three of the cases are illustrated in Higure AS. Each case shows life cycle costs (at a 12 percent discount) for the first 14 of the 31 maintenance alternatives for paved roads specified in Table Al. In case A, with average daily traffic (ADT) of only 500 veheles per day and 11Sht axle loadings, it would suffice to use a normal. design (82.0) comesnsurate with the estimated ESAL since .the life-cycle costs are consistently higher for a higher strength compensat- Ing design (8043.3), under all maitenance assumptions. But with heavier traffic a different conclusion emerges, as shown by Case C (ADTo2,500 and heavy axle loading). In this case, a normal full-strength pavement (Wis) will have a lower life-cycle cost, under almost any assumption about future maintenance, than one with a lower initial strength (83.3.5), which repre- sents the time-staging option. At or above this combination of traffic and axle lodgings it would not pay to consider the time-staging option. 23. letween these two limits, the decision for or against time-staging of construction is reacbhd by balancing the potential loss If no maintenance will be done againt the cost saving from dme-staging that would be realised If future maintennce (including strengthealng) io performed as desired. In this way a thresbold probability of performing good maintenance In the future is obtained, above which the time staging option could be justified. Tor Case B In Figure AS, for example, with an ADT of 1,000 vpd and light axle 51 A normal full strength pavement is defined as one deslgned on the basis of accepted pavement engineering principles to carry a specific number of cumulative Equivalent Standard Axle toads (ESAL), until an unacceptable levl of functlonal serviceability Is reached. To compensate for inadequate maintenance, a higher pavemet strength than given by normal deslgns may be warranted In certain situations. 6/ 1Tie-staglng at the network level is effective only hben the condition of pavements is regularly monltored and evaluated with an appropriate pavement management system. Othewise It is difficult to predict the critical points In the pavemnt life when a major maintenau Intervention may be lnded to prevent premature structural failure0 3' Cm (**CmA 9*35 _TMIS 0.04 million ESAL per year per lane X~ ~ ~~A I ^s _ ~ _ ' CAS Cm(WON. 100. i$. CAN x3 .21 .UU § \ ~~~~~~~~t_e~ 0.05 mil ESAL v 3\ per year per lne ~- I *33c t 5 C C?-I 2-1 VS L costs$iscounted at 12'Pe"t a eeyr prl ii 4 Figture AS Inlunc of Iitnae Policies o Seetio of Intia Pavent Strent (based orn present alue of life cycle costs9 disconted at 12 percent a yea) - 72 - loads, thls threshold probability Is estimated at 33 percent.7/ In this case, tlie staging of pavement construetion should only be coasidered If there is a one-third or better chance of good ctritenance being performed In the future. 24. Under fundlg arrangements that favor constructton over maintenance or where exterrnl aid agencies are willing to finance construction but not maintenance expenditures, It is often expedient to forgo maintenance until It becomes necessary to reconstruct the pavmenit. ALT3 In Case B codifies such a asintenance stategy, an option that would also reduce life-cycle costs significantly, It there Is fair expectatlon that futds for reconstructWou would be forthcoming when needed. In this case, a less than full-strength 4esign could be considered for Initial construction, If the probability of obtainlig future funds for reconstruction Is 45 percent or better. 25. Table A3 Illustrates how the decislon may varf In a wider array of cases, It maps out combinatlons of discount rate, dally traffic, and axle loading where the decision for or against time-staging io Independent of the degree of uncertainty about future maintenance. In a middle area, however, the reliability of future malntenance matters. Timw-staging is thus general- ly to be preferred If the probability of adequate maintenance In the future exceds 30 or 75 percent, depending on traffic volumes and light aile loads. At higher discount rates (in this case 24 percent) tlmestaging tends to be a preferred choice even vith high traffic volumes and axle loading and a rela- tively low probabllity of future maintenace. Ubere capital is cheap or a greater weight is attached to long-term benefits (an lntergenerational Issue), stronger nlitial pavements based on normal design generally dominate (say, at a discount rate of 6 percent). couoiema Traffic Thres 'lds for Maima gravel loeds 26. Even with the best malatenance practice, vehicle operatlig costs on gravel roads are between 10 and 30 percent higher than on paved roads. In addltion, the present value of the cost of routine malatenance and resur- facing Is between five and eight time greater for a well-maintained, high- volume gravel road than for a newly built paved road. Paving is therefore 71 This may be derlved as follows for Case B: Let x - probability of optimal maintenance (ALT 14) 1-z a probability of ail maintenance (ALT 00) Dopto difference In life-cycle costs for Initial pavement strength of 9N 3.5 and SN 2.0; under ALT 14. Dni- differeane In llfe-cycle costs for Initial pavement design strength of S8 2.0 and SN 3.5, under ALT 00. Then, from the inequalityt (Dopt).x > (Dnl)(1-x) or (1.27-1.23)x>(1.33-1.31)(1-') Therefore, x>0.33 - 73 - Table U3 Criteria for Selecting Time-StagiBn Optiton for Pavement Desin,1 8 Minimum Maintenance Reliability for Time-Staging Option b/ S Discount late of ADT Loadig c/ 61 12X 24X Compensatory design not applicable; low 300 light tO mdeium strength paveents heavy 30% (SN < 3.5) based on normal design, adequate for thls 500 light 30X level of heavy n.a4/ 752 traffic 1000 light 9SS 30X loading. heavy soa.x/ 80. 2000 light 9SX so15S heavy 90S 15X Time-staging not 2500 light applicable; high 952 502 heavy strength pave- 602 meats (SN > 4.0) based on normal full-strength desigp appropriate. aT She economic costs for paving (in thousand US$ per kn, In 1964 prices, assuming all earthworks and structures already In place, were estlmated ass 8N Costa Riea ml 2.0 19.6 27.2 3.5 68.5 59.3 5.0 118.8 129.4 b/ nlamum probability of adequate maintenance In the future; time staging conditional to this malium level of mainteance reliability, otherwis use normal strength design. C/ Light axle loading Is representative of conditions In Costa Rica (0.05-0*10 million SAL p.a. per lane); heavy axle loading is representative of conditions In NBal (1.2 millon NMAL peso per lane). d/ n.a..time staging optlon not applicable; use normal strength design. - 74 - Indicated iuhen the expected svinpgs In vehicle operating and maintenace costs (relative to a well-malntalned gravel road) exceed the preseat value of the paving cost. 27. These cost tradeoffs now seen to cover a wider range of traffic volumes than had commonly been assumed. Applications of the Bank's model In Costa Rlca and Mall Indicate that the breakeven traffic volume for pavitig may vary from under 100 vehicles a day to more than 400, depending on the costs of paving, the discount rate, the rate of growth of traffic, and the anticipated quality of future maintenance. 28. Paving thresholds are also sensitive to variations In assumptlons about the quality of maintenance. Where experience shows that the probabil- ity of adequate future maintenance is low, the traffic threshold for paving is lowered. And If there are no budget constraints, an economic case can be made for an all-weather paved road that may remain almost free of major main- tenance Interventions for seven to ten years. In looking at total costs, however, more frequent regraveling and grading operations carried out effi- ciently are likely to prove more economical than paving, particularly If the future availability of maintenance funding is uncertaln. 29. Paving gravel-roads In arid sones Is sonetimes suggested as a means to alleviate the discomfort and Inconvenience of travel oh dusty roads, but It is difflcult to quantity these benefits. Comon observation suggests that vehicle speeds and passing opportunities on such roads are severely restrict- ed by reduced visibility. The result is a traffic safety hazard similar to fog and congestion akin to that caused by heavily loaded slow-moving vehicles on narrow roads. Not enough empirical Information Is available on traffic flow and vehicle operatlng characteristlcs in a dusty environment to evaluate the benefits of paving roads In arid or desert areas. 30. Lower paving thresholds may also be Indicated for roads located in river deltas (lower Iangladesh), old lake beds, sandy deserts, and low coastal areas due to the scarcity of gravel deposits and other sources of aggregate. In mountainous terrain, excessive maintenance costs resulting from erosion caused by rapid runoff, and frequent replacemeat of gravel, may warrant protecting the road surface with a bituminous surface. A possible technical alternative to using a gravel surface in these situations is to .stabilize in-situ soils by small amounts of a suitable binder (bitumen, cemeat, lime, or fly ash) and then to protect the stabilized material from weathering and traffic, where necessary, by a light bituminous seal. The IMesst of Ouerloadimu 31. The onerous impact of overloading on pavement performance is reflected in the progression of roughness over tlme, given different maintenance policies and loading conditions on-otherwise similar pavements (Figure A6). Case A is typical of Mali - 33 percent trucks with heavy axle loads, accumulating to about 14 million SAL In 25 years. Case I is typical of Costa Ric - 25 percent trucks with relatively light axle loads, * 9 IL a I.' U js '9 I '9 U - 1 3. '9 1 Ii U I t04 '9 I. '9 1 I 3 6 '9 I I .4 '9 1* I '9 1 I S I S .4 '9'9 e '9 S S 5 '9 'I' I I 6% '9 15 11 I 0 4J 6. 6 164 .4 N 11h1 6 *0  - 76 - accumulating to about 2.5 million ESAL per lane over 25 years.81 Heavy loading signlficantly accelerates the rate of deterioration, ad ia this comparlson the need for overlays in Case A is advanced by some six years. Even with optimal maintenance policies, the average life-cycle roughness remains higher with heavier axle loads. The results for Case A are typical of what has happened to pavements designed on the premise that axle load lImits would be enforced, often an unrealistic assumption. 32. Although empirical evidence shows that a major part of pavement damage increases with the fourth power of the axle load, the regulation of axle loads has proved exceedingly difficult. Many road authoritles have therefore built stronger and more expensive pavements based on actual pave- ment loadings than would be necessary with adequate control of axle loads. Expected changes In road transport technology resultlig from road lmprove- mets (such as the Introduction of combination vehicles In place of single unit trucks) should be taken into consideration in pavement design. 33. Research on the economic optium for axle loads suggests that, limits In the longer term should be increased beyond the prevailing 8-10 ton single (and 13-16 ton tandem) axle loads; the economic gain from lower unit transport costs outweighs the loss In lncresed road da&Age. Zn the short term, bowever, the existing roads and bridges are not strong enough to carry heavler loads. So, a compre'enssive plan for bridge and pavement strength- ening should accompany an lncrease in axle load lImits. 8/ The actual axle loads, on average, are about 60 percent beavler io Case A, partly due to the higher axle load limits (13-ton single axle) In MNal and partly due to overlooding. 77 - Short bibliography and background papers 1. AASHTO, A Manual on User Benefit Analysis of Highway and Bus Transit improvementse (American Association of State Hlghway and Transportation Officials, Washington, D.C. 1977). 2. Baker, M., Rebuilding America's Infrastructure - Agenda for the 1980'8. (Duke Press Policy Studies, Durhas, N.C., 1984) 3. Bhandari, A., and K. Sinha, Opti ml TVImng for Pavlg Low Volume Gravel Roads. Transportation Research Board, TRP No. 702. Low Volume Roads; Second International Conference, Ames, Iowa 1979. 4. Bhandari, Anill, et al., Road Deterioration ln Developing Countries: Technical options for Road Mantenance and Economic Consequences (Paper presented at the TM Annual Meetings, Washington * D.C.,- January 1987). 5. Buchanan, James A., *he Economics of Earmarked Taxes.' Journal of political 1Eo n o, (October 1963). 6. Cheasher, Andrew, and Robert Harrison, Vehicle Operatlag Costs: Evidence from Developi Countries. (fcrthcoming publlcatlon of the World Bank In the DUm series). 7. Choate, Bad Roads - The Hidden Costa of Neglect, (National Asphalt Association, RLiverdale, Maryland 1983). 8. Coukis, Basill, et al., Labor-Based Construction Progrm_s. A Practical Guide for Plauning and Management. (World Bank, Washington1 D.C. 1983). 9. Cox, Brian B., Evaluation of Incentives for Efficency in Road Maintenance Organization: The U EK. Eperince. (Transportation Department Discuscion Paper, Transportation Issues Series, Report No. TIP4, World Bank, Washington, D.C. 1987). 10. de Weille, J., Quantification of Road User Savings, World Bank Occasional Paper No. 4, 1966). 11. Fais, Asif, et al., Road Deterioration bei.lopinq Countries: The State of the Road Networks and a Country T lop of Resonse Meaures, (Paper presented at the TD Annual Meeting, Wahington, D.C., January 1987). 12. Harral, C., and J. Eaton, "Improving Equipernt Management in Highw Authoriltes In Developitg Countrlos." (Transportation Department Discussion Paper, Transportation rssues Series, Report No* TBP2 - World Bank, Washington, D.C. 1986). 13. Harral Clell, et al., Evaluating the Economic Priority of Highway Maintenance. (Paper presented at the Senior Management Semnaar oan Planning, Fnnanclng and Managing Cost-Effective Road Mantenance, Abidjan, June 1985). - 78 - 14. Harral, C., World Bank Experience with Highway Maintenance: Tlme for New Directions? ~Paper presented to the PTRC Annuar Meeting, London, July 1985). 15. Harral, Clell, et al., An Appraisal of Highway Maintenance by Contract ln Developi g Countrieso (Transportation Department Discussion Paper, Transportation Issues Series, Report No. TRPI, World Bank, Washington, D.C. 1986). 16. Harral, C., Road Deterioration in Developing Countriest Organization and anagement of Road Maintenance. (Paper presented at TRB Annual Meeting, Washington, D.C., January 1987). 17. &atry, et al., How Effective Are Your Colmmity Servicos? Procedures fot~~ Of Nuninx for Monitor!B, the Effectiveness o nicipal Services. (We' Urban- Institute Washington, D.C. 1977). 18. IRP, World Road Statistics, 1980-84; (International Road Federation, Washington, D.C. 1985). 19. International Road Transport Union, World Transport Data, (Geneva 1985). 20* Hicks, Norman, and Ann Kablech, Effects of lxPendlture Reductions in Developlig Countries. (CPD Working Paper, November 10, 1983). 21. Hirschmans, Albert, Exit, Voice, and Loyalty, (Harvard University Press, 1970). 22. Johansen, Frida, Deterioration of the Road Networks - an Emerging Crisis. (TRP Departmental Paper, 1984). 23. Mason, N., Road Maintenance Survey for Meot and Central African Countries. (Paper presented to West and Central Africa Seminar on Cost-Effectlve Road Maintenance, Abidjan, June 198S.) 24. Mason, ., and S. Miquel, Results of Road Maintenance Survey for Latin American and Caribbean Countries, (World Bank, Washington, D.C. 1986). 25. Oakland, William R., Earmarking and Decentralizatlon. (Paper to the 77th Annual Conference on Taxation, Nashville, Tenn., November 1984). 26. -, A Survey of Arguments and Practice In Tarmxrking Taxes. (Tulane University, 1985.) 27. Paterson, William D.0., Road Deterioration and Maiatenance Effects: Models for Planninag and Magement. (forthcoming publication of the World Bank ln the HD4 series). - 28. Ray, A., et al., The World Development Report, 1986, The World Bank, Washington, D.C. 1986. - 79 - 29. Robinson, R., et al., The Road Maintenance Crisis in Africa, An Asenda for Action. Paper presented at the Senior Management Seminar on Planning, Financing and Managing Cost-Effective Road Maintenance (Abidjan, June 1985). 30. Robinson, R., Road Maintenance - Planning and Management for Developing Countries. (Overseas Unit, TRIL 1985). 31. Robinson, R., "Maintenance Management for District Engineers," Overseas Road Note 1. U.K. Transport and Road Research Laboratory (1981). 32. Rolt, J., Optisum Axle Loads of Commercial Vehicles in Developing Countries. (TRRL Laboratory Report 1002, Crowthorne, Berks., 1981.) 33. Sayers, Michael W., et al., The International Road Roughness Experiment: Establishin% Correlation and a Calibration Standard for Measurements. (World Bank Technical Paper No. 45, 1986). 34. Sayers, Michael W., et al., Guidelines for Conducting and Callbratlng Road Roughness Measurements* (World Bank Technical Paper No. 46, 1986). 35. Smitht G., and C. Harral, Road Deterioration in Developing Countries: Financial Requirements. (Paper presented at TRB Annual Meeting, Washington, D.C., January 1987). 36. Transportation Research Board, Consequences of Deferred Maintenance. (Synthesis of Highway Practice No. 58, TRB, Washington, D.C. 1979). 37. UK Department of Transport, Policy for Roads; England 1980. (H8Os, London, 1980) 30. 38. U.K. National Development Office, Investment in the Public Sector Built Infrastructure - Report A: Roads and Bridget, p.9, London, 1985. 39. United Nations, Annual Bulletin of Transport Statistics for Europe, New York, 1983. 40. U.S. Congressional Budget Office, Public Works Infrastructure - Policy Consideration for the 1980s. (U.S. Congress, Washington, D.C. 1983). 41. Watanatada, Thawat, et al., Vehicle Speeds and Operating Costs: Models for Road Planning and Management. (forthcoming publication of the World Bank in the HDM series). 42. - , The Highway Design and Maintenance Model: Volume 1, 1DM-Ill Model Description. Volume 2, HDM-Ill User's Manual (forthcoming publications of the World Bank in the 1DM series). 43. Winfrey, R., Economic Analysis for Hilghway (International Textbook Company, 1968). 44. The World Bank, The World Bank Atlas, 1986, 19th Edition, Washington, D.C., 1986. s8o - 45. Benor, D., et al., Agriculture ExteAsion - The Training and Visit System, (World Bank, Washington, D.C. 1984). 46. Trataportation Research Board, Technology Transfer to Tran gprtatlon Agencies, (Record No. 1080 TRB, Washlngton, D.C. 1986). 47. Cooper, Lauren, The Twinning of Institutions - Its Use as a Technial Assistance Delivery 8ystem. Technical Paper No. , (World Bank, Washington, D.C. 1984). 48. U.S. Department of Transportation, America's Eighways 1776 - 1976, Federal Highway Administration (Washligton, D.C. 1976).