Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized

,
17
A WORLDBAM,1itv

7
13370
POLICYSTUDY

Causes and Remedies

RoadDeteriorationin DevelopingCountries
Road Deterioration
in Developing Countries

A World Bank Policy Study

I
Road Deterioration
in Developing Countries

Causes and Remedies

The World Bank

W'ashington,D.C.
Road Deterioration in Developing Countries: Causes and
Remedies was prepared under the direction of Louis Y.
Pouliquen, director, Infrastructure and Urban Development
Department. Clell Harral and Asif Faiz were the principal
authors. Substantial contributions were made by Esra
Bennathan, Graham Smith (chapter 4), and Anil Bhandari
(chapter 2). Frida Johansen made an initial assessment of
the global magnitude of road deterioration. Per Fossberg
and William Paterson provided help on technical and engi-
neering issues, and Edward Holland assisted in drafting the
report. The guidance and comments given by S. Shahid
Husain, Parvez Hasan, Curt Carnemark, Sir Alan Walters,
and Gregory Ingram are gratefully acknowledged. Thanks
also go to regional transport staff who reviewed the report
at various stages; Meta de Coquereaumont and Bruce Ross-
Larson for editing; Rodrigo Archondo-Callao and Olivier
Bottrie for research assistance; and Mari Dhokai, Marjeana
Gutrick, and Pamela Cook for typing.

Copyright )1988 by The International Bank

for Reconstruction and Development / The World Bank
1818 H Street, N.W., Washington, D.C. 20433, U.S.A.

All rights reserved

Manufactured in the United States of America
First printing June 1988

The judgments expressed in this study do not necessarily

reflect the views of the World Bank's Board of Executive
Directors or of the governments that they represent.

Library of Congress Cataloging-in-PublicationData

Harral, Clell G.
Road deterioration in developing countries: causes and
remedies.
p. cm.-(A World Bank policy study, ISSN 0258-2120)
"Prepared by Clell Harral and Asif Faiz, with
contributions by Esra Bennathan, Graham Smith, and
Anil Bhandari"-T.p. verso.
Bibliography: p.
ISBN 0-8213-1039-9
1. Roads-Developing countries-Maintenance and
repair. 2. Roads-Developing countries-Deterioration.
1. Faiz, Asif. 11. International Bank for Reconstruction and
Development. 111.Title. IV. Series.
TE220.H37 1988
388.1'109172'4-dcl9 88-9663
CIP
 Foreword

The failure to maintain roads is tantamount to an act Road deterioration is not confined to developing
of disinvestment, for it implies the sacrifice of past countries. Nor is it a new concern for the World Bank,
investments in roads. Over the past two decades an which published The Road Maintenance Problemand
estimated $45 billion worth of road infrastructure has International Assistance in 1981to draw attention to
been lost owing to inadequate maintenance in the the matter. What is new, however, is the scaleto which
eighty-fivedevelopingcountries reviewed in this poli- road deterioration has progressed in so many develop-
cy study. This loss could have been averted with ing countries. In this study we have attempted to
preventive maintenance costing less than $12 billion. estimate the physical and financial magnitude of the
The loss of physical infrastructure is only part of deterioration and to identify remedial measures ap-
the picture. Bad roads seldom deter users or curb the propriate to the circumstancesof different countries.
volume of traffic. Instead, they raise the cost of road We have also tried to determine the principal causes
transport-the dominant mode of transport for both of road deterioration and the reasonsthe problem has
people and freight in most countries. A dollar reduc- become so widespread. Economic adversity is part of
tion in road maintenance expenditures can increase the explanation, past mistakes in investment choices
the cost of vehicleoperation by two to three dollars. are another, but a large part of the problem has to
Insufficientspendingfor maintenance thus exacts hid- be attributed to institutional failure in the countries
den costs several times the cost of maintaining and themselves.
restoring roads. Road users bear the brunt of these The quantification of losses underscores the im-
additional costs, which dwarf the savings to a road portance of the problem, and the search for causes
agency from deferring or neglectingmaintenance. points the way to needed change and adjustment.
Much of the problem of road maintenance is rooted Fortunately, the magnitude of the task is lessened by
in its economic and institutional aspects. Inadequate advances in engineering knowledge and experience
incentives and weak accountability derive from the that permit greater flexibility in the design of road
characteristic separation of responsibilities and con- maintenance strategies and a wider selection of solu-
trol between the providers and users of roads. Unlike tions. This study describes some of the options and
most other types of infrastructure, roads are neither presents an institutional framework for implement-
built nor maintainedby those who use them to market ing them in developing countries. It provides an
output or services. estimate of the resources needed to remedy the situ-

V
ation and suggests ways to use these resources ef- veloping countries. Without corrective action, many
ficiently. The recommendations are addressed to the more developing countries may soon face a situation
developing countries, the lending and donor agen- in which poor roads, and therefore inadequate road
cies, and the development community at large, in- transport, become an insurmountable obstacle to
cluding the World Bank. economic recovery and growth.
This report provides an early warning of a trans-
portation crisis in developing countries that will W. David Hopper
occur if the escalating damage to roads is not con- Senior Vice President
tained. The portents of such a crisis are visible in the Policy, Planning, and Research
deteriorating road transport conditions in some de- June 1988 The World Bank

vi
 Contents

Glossary ix
Road Conditions ix
Road Maintenance and Improvement Works ix
Acronyms and Initials x
Data Notes x

Summary 1
The Repair Billlfor the Next Ten Years:
$90 Billion or Much More? 1
What Caused the Deterioration? 1
The Requirements for Efficient Restoration 2
Implications for the World Bank 3

1. The State of the Roads 5

An Overview of Road Conditions 5
Major Determinants of Road Conditions 7
The Hard Core of the Problem 9

2. Technical Options and Their Economic Consequences 11

The Fundamental Relations 11
Strategies for Road Investment and Maintenance 12
Tactics under Budgetary Constraints 15

3. The Institutional Challenge 17

The Constraints 17
The Search for Solutions 18
Elements of Reform 19
Technical Assistance 21

4. Financial Requirements 23
The Overall Picture 23

vii
 Country Differences 25
Marshaling Domestic Resources 27
External Financing and Assistance 28

5. Conclusions and Policy Recommendations 31

World Bank Policy 32
Action by the International Community 34

Appendix A. Statistical Tables 37

A-1. Basic Characteristics of Road Networks by Region 39
A-2. Road Networks by Country and Region, 1984 40

Appendix B. Exploring Cost-Effective Options for Road Investment and Maintenance 45

Bibliography 59

Boxes
1-1. The Consequences of Road Neglect in Ghana 6
1-2. Nigeria's New Roads and the Risk of Massive Deterioration 8
1-3. A Case of Undermaintenance: Brazil's Federal Highway Network 9
1-4. The Construction-Maintenance Tradeoff: Consequences for the Road Network 10
2-1. The Highway Design and Maintenance Standards Study 13
3-1. Overstaffing and Resource Imbalances in Kenya 18
3-2. Underutilization of Equipment in Western Africa and Latin America 20
3-3. Twinning of Turkish and U.S. Highway Organizations 21
4-1. Earmarked Road Funds in the Central African Republic 28
4-2. World Bank Lending in Chile 29
5-1. A Diagnostic Framework for Determining External Assistance Policy in the Road Sector 33

viii
 Glossary

Road Conditions structural integrity and ride quality. A paved road

normally needs resurfacing at the transition from
Good. Paved roads substantially free of defects and good to fair condition, provided the volume of traffic
requiring only routine maintenance. Unpaved roads justifies retaining it in good condition. Resurfacing
needing only routine grading and spot repairs. is sometimes called "periodic maintenance," even
though all maintenance activities are periodic. Costs
Fair. Paved roads having significant defects and re- can vary from less than $8,000 a kilometer to more
quiring resurfacing or strengthening. Unpaved roads than $40,000a kilometer.
needing reshaping or resurfacing (regraveling) and
spot repair of drainage. Rehabilitation. Selective repair, strengthening, and
shape correction of pavement or roadway (includ-
Poor. Paved roads with extensive defects and requir- ing minor drainage improvements) to restore struc-
ing immediate r ehabilitation or reconstruction. Un- tural strength and ride quality. The term "strength-
paved roads needing reconstruction and major drain- ening" is sometimes used to describe a specific
age works. category of pavement rehabilitation involving the
application of overlays. Costs of rehabilitation can
Road Maintenanceand Improvement Works vary from less than $30,000 a kilometer for un-
paved roads to more than $200,000 a kilometer for
Routine maintenance. Local repair of roadway and paved roads. The costs for paved roads, however,
pavement; grading of unpaved surfaces and shoul- rise steeply as a pavement deteriorates from fair to
ders; regular maintenance of road drainage, side poor condition.
slopes, verges, traffic control devices, and furni-
ture; roadside cleaning, dust and vegetation con- Reconstruction. Renewing the road structure, gen-
trol, snow or sand removal, and maintaining rest erally usingexisting earthworks and road alignments,
areas and safety appurtenances. Typical costs range to remedy the consequences of prolonged neglect or
from less than $300 a kilometer to more than $5,000 where rehabilitation is no longer possible. Costs can
a kilometer. vary considerably, ranging from about $45,000 a
kilometer to more than $300,000a kilometer.
Resurfacing. Regraveling an unpaved road or resur-
facing a paved road (with a thin asphalt overlay, a Restoration. Major rehabilitation and reconstruction
surface treatment, or a seal coat) to preserve its works considered together.

ix
Betterment. Road improvements related to the width, Acronyms and Initials
alignment, curvature, or gradient of road (including
associated resurfacing and rehabilitation works) to IDA International Development Bank
improve traffic speed, safety, or capacity. Betterment IBRD International Bank for Reconstruction and
works are not considered maintenance activities ex- Development
cept for ancillary road rehabilitation or resurfacing GNP gross national product
operations. Costs can vary considerably, depending IRI International Roughness Index
on the geometric improvements. HDM Highway Design and Maintenance Stan-
dard3 Study
New construction. Constructing a paved, gravel, or HDM-III Highway Design and Maintenance Stan-
earth road on a new alignment; upgrading a gravel dards Model
or earth road to paved standards; providing addi- PRA Public Roads Administration
tional lane capacity; or constructing additional car- Data Notes
riageways, frontage roads, grade-separated inter-
changes, or multilane divided highways. Costs of "Dollars" are current U.S. dollars unless otherwise
new construction can vary from less than $50,000 a noted. "Billion" is 1,000 million. "Tons" are metric
kilometer for a gravel road to more than $1 million tons, equal to 1,000 kilograms, or 2,204.6 pounds.
a kilometer for a four-lane access-controlled divided Numbers in tables may not add up to given totals
highway. because of rounding.

x
 Summary

The developing countries have lost precious infra- In addition, another 40 percent of the paved roads
structure worth billions of dollars through the dete- in these countries require strengthening today or will
rioration of their roads. If they do not immediately in the next few years. This work, along with routine
begin to do much more to preserve their roads, they maintenance, will cost another $40 billion to $45
will lose billions more. Large road networks, built at billion over the next ten years. That brings the total
great expense, have been undermaintained and more bill for these countries to about $90 billion. But if
heavily used and abused than expected. If this neglect these countries do not improve their management of
continues, the deterioration of roads will accelerate roads, the eventual cost of restoration could easily
as old pavements crumble and new ones outlive the increase two- to threefold, and the cost to users
initial period during which the effects of neglect are would rise even more.
barely perceptible.
The cost of restoring these deteriorated roads is What Caused the Deterioration?
going to be three to five times greater than the bill
would have been for timely and effective mainte- In the 1960s ancL 1970s road networks expanded
nance and restoration is only part of the cost. Vehi- much faster than the corresponding maintenance
cle operating costs rapidly outpace the costs of road budgets and institutional capacities. Traffic has also
repair as the condition of roads passes from good to become much heavier than expected, and axle load-
fair to poor. Together, these avoidable costs are going ings have often exceeded the designed capacity of
to form a formidable obstacle to further economic pavements. These patterns are evident almost every-
development. where. Less evident is what has been happening to
the roads.
The Repair Bill for the Next Ten Years: New paved roads, if inadequately maintained,
$90 Billion or Much More? deteriorate slowly and almost imperceptibly during
the first half to two-thirds of their service life, de-
In the eighty-five countries that have received World pending on the traffic. After that grace period, which
Bank assistance for roads, a quarter of the paved may last ten to fifteen years, the pavements deterio-
roads outside urban areas need reconstruction-as rate much more rapidly. Without timely maintenance
do a third of the unpaved roads. This work will cost they break apart.
$40 billion to $45 billion. (Timely preventive mea- As roads become rougher, the costs of operating
sures costing less than $12 billion could have saved vehicles-and of transporting goods-begin to shoot
these roads and held down the operating costs for up. The neglect of maintenance continues, however,
road users.) because it is the vehicle operators that pay these

1
costs-and pass them on if they can. Road authori- and maintenance strategies must therefore be tailored
ties are not directly affected by these costs, and they to the circumstances of individual countries. Nev-
come under no immediate pressure to improve road ertheless, some general guidance can be offered in
conditions. Road users are often slow to see the link light of our growing understanding of the causes,
between road conditions and the prices of goods and process, and effects of road deterioration.
transport services-and are usually not organized to For one thing, the savings in vehicle operating
do something about it. costs from paving lightly traveled roads are less than
In the absence of public pressure and lacking a previously estimated. Furthermore, the total life-cy-
clear understanding of the seriousness of the prob- cle transport costs on paved and unpaved roads are
lem, few governments have given road maintenance nearly the same over a wide range of traffic volumes
a high priority in their budgets. The urgency of the (typically 150-400 vehicles a day), as long as the
situation has not always been fully appreciated by roads are reasonably maintained. Big differences in
all donors and lending agencies, some having been cost arise only if maintenance is not done or is
readier to provide funds for new construction than deferred. If the availability of funds for future upkeep
for maintenance and restoration. New construction of paved roads is uncertain, it is economically safer
has sharp political visibility,maintenance little glam- to keep lightly traveled roads unpaved and reason-
our. Inadequate maintenance in developing countries ably maintained.
has various causes, but institutional failure is the only It also makes sense-in times of budgetary strin-
explanation for its wide extent. At the heart of this gency when road allocations are not too far below
failure is the absence of public accountability. the optimum-to consider alternative maintenance
strategies. Some strategies that involve small cuts
The Requirements for EfficientRestoration in spending barely affect the costs to road users.
Others, although similar in amount, can drive up
The task now facing many developing countries-as user costs substantially. Still other strategies, in-
well as the World Bank and other aid agencies-is to volving much deeper cuts, can multiply the future
salvage roads that have deteriorated severely and to cost to both the road agency and the user-so
protect newer roads from a similar fate. This task has much that they are self-destructive. For example, if
large financial, technical, and institutional require- the budget cuts are so large that they preclude the
ments. Of these, the institutional requirements are resurfacing or strengthening of the paved roads
the most pressing. that are now in fair condition, these roads will
The technical options today seem richer. Thanks soon fail structurally and require much more costly
to recent empirical research that has broadened our restoration. This, regrettably, is the situation in
understanding of the physical process of road deteri- many developing countries, and it will soon spread
oration, we now know of ways to keep roads service- to many more.
able at lower cost than before. The financial re- Here are some of the technical choices that each
quirements, however, are steep. Some countries will country should examine in light of its needs and
require large infusions of external capital; all will capabilities:
need political resolve to mobilize resources at home. * Gravel roads should be paved only after a complete
But additional financial resources alone will not analysis of the costs, climate, present and future
ensure effective restoration and preservation of traffic, and reliability of future maintenance.
roads, nor will they prevent a recurrence of the kind - If the volume of traffic indicates that paving is
of crisis and waste so evident today. Needed above advisable, the choice of pavement strength should
all is a reform of the institutional base of the road be guided by the likelihood of undermaintenance
sector. Accountability must be stronger in all entities and excessiveaxle loads. If the likelihood is high,
and activities that maintain public assets. Public and pavements should be built to the required strength
political attention must be deliberately fostered. The immediately rather than in stages. The cost will,
organization, manning, and activities of the institu- of course, be higher, so that a given budget for new
tions looking after roads must be improved to in- construction will yield fewer kilometers of new
crease their absorptive capacity. road.
* Deterioration of paved roads is gradual and barely
TechnicalRequirements perceptible during a long initial phase that lasts
up to two-thirds of the pavement's life cycle.
Standard engineering practices have very different Resurfacing and strengthening can thus be de-
effects in different environments. Road investment ferred somewhat-as long as the roads remain in

2
 fair condition and have not entered the critical FinancialRequirements
phase of their life cycle. Longer deferral will cause
pavements to break up. The stage of a paved road The present bill for repair and maintenance is about
in its life cycle must therefore determine the $90 billion, or $9 billion a year for the next ten years.
decision to defer resurfacing and strengthening That could arrest future deterioration and clear up
operations. the backlog of economically warranted restoration.
Traffic is critical for maintenance decisions. If If the work were to be finished in fiveyears, however,
funds are tight, it may be best to keep heavily the total cost, again without discounting, would
traveled roads in fair or good condition and to come down to less than $70 billion, or about $13.5
reduce substantially the maintenance of roads with billion a year. (Road spending in 1984 was about $13
light traffic. Even if funds are not tight, it may billion, but only half of this went for restoration and
make little economic sense to fully restore paved maintenance.)
roads that are lightly traveled. If funds are very These aggregate sums obscure wide differences
tight, it may be best to let lightly traveled roads from country to country. The sixty-one developing
deteriorate further and perform only minimal rou- countries that have data on road spending fall into
tine maintenance to keep them usable. three categories of capability-with about a third in
each. First are the countries that could meet future
InstitutionalRequirements maintenance requirements and eliminate their back-
logs in five years without increasing their total spend-
The inadequacies of road maintenance stem in part ing on roads, but only by holding back on new
from the structure and functions of the traditional construction and allocating up to 80 percent of their
road agency. Often a public monopoly, the agency total spending to restoration and maintenance. Next
has too many responsibilities-for planning, control- are the countries that could do the same in ten years,
ling, and executing both construction and mainte- but only by raising their spending for those years by
nance. The agency typically devotes too many staff, 50 percent and by allocating 80 percent of that total
funds, and facilities to execution-to the detriment to restoration and maintenance. Last are the coun-
of planning, control, and evaluation. This makes it tries that would have at least to double their spend-
desirable in many countries to separate the execution ing, even if they devote all of it to restoration and
of works from the other functions by lodging it in the maintenance.
private sector or in a separate government agency. To justify more spending on roads and its reallo-
Where this separation has succeeded, the incentives cation to restoration and maintenance, each country
for good performance have been strong and the will have to perform a systematic analysis to deter-
delineation of accountability clear. mine how much should be spent on roads-and how.
A road agency also needs an effectivemanagement Some countries may choose to revise user charges or
information system to monitor traffic, road condi- impose new ones, which if translated into better
tions, and (just as important) its internal workings. roads can reduce rather than increase vehicle operat-
The agency shoulcLbe equipped to analyze the life- ing costs. Some may decide that to avoid misalloca-
cycle costs of construction and maintenance and the tions they will have to earmark budgeted funds for
present and future cost to users. Testing alternative restoration and maintenance, especially the funds
design and maintenance policies for their sensitivity that might be raised through a temporary surcharge.
to different discount rates, traffic loads, and other Others will have to rely heavily on external sources
variables may be helped by computer models. But of finance.
even less sophisticated methods to plan investments
and maintenance will be unworkable without a reli- Implications for the WorldBank
able data base and capable staff.
The road agency should, above all, introduce me- Economic losses-actual or potential-caused by in-
chanisms to increase the internal accountability for adequate road maintenance can have important im-
performance and for resource use-and to provide plications for the Bank's macroeconomic dialogue
incentives for goocl performance in the agency and with its borrowers, particularly in the context of
by the contractors it engages. It should, moreover, public expenditure priorities. The evidence in this
work with the media and with nongovernmental study suggests that in several countries the economic
organizations to make the political leadership and returns from spending more for road maintenance
the public aware of the effects and high costs of would justify a substantial expansion of such expen-
failing to maintain the roads. ditures relative to outlays for new road construction.

3
An expansion in total expenditures on roads relative and of the high cost of neglecting roads.
to other sectors could also be warranted in some The Bank will also promote more coordinated
countries. That determination, however, can be made action by the international community in dealing
only by a comprehensive review of public expendi- with the growing problem of road deterioration in
tures which evaluates the comparative benefits of the developing countries. To this end the Bank will
proposed outlays in all major sectors. work closely with other aid agencies in designing
The developing countries differ according to their road programs appropriate to the needs of each
need to expand and improve their road maintenance country (along the lines of the four categories spelled
organizations and procedures, their need to increase out above). The Bank will also work closely with
their total road budgets and make reallocations from other aid agencies in supporting research on roads-
construction to maintenance, and their need for ex- work that would be done mainly by road agen-
ternal assistance. In addition, some have obsolete cies and road research institutions in the develop-
networks, others new ones. These differencessuggest ing countries. The emphasis in these efforts will be
several distinct categories of need for external assis- on international exchanges of data, technology, and
tance in the road sector. The Bank's lending for roads management information systems. The Bank will
should be tailored to the circumstances in each cate- support compilation and publication of statistics
gory-and be made conditional on an acceptable on roads-work that should be undertaken by the
distribution of road expenditures for maintenance, United Nations Statistical Office or an organization
restoration, and new construction. such as the Permanent International Association of
* For countries with the best performance (category Road Congresses or the International Road Feder-
I in the text)-largely but not exclusively upper- ation. External agencies should provide financial
middle-income countries-the Bank will view it- assistance for this effort.
self mainly as a vehicle for institutional and tech- The recommendations for World Bank policy
nological improvements. It will nevertheless ex- emerge from a view of the root causes of road
tend funds for balanced programs of maintenance deterioration in countries where it is now severe.We
and new construction. know now that some past investments have been
* For countries in which road maintenance is inade- mistakes. We know, too, that deterioration revealsits
quately funded (category II)-including a substan- symptoms late and gives little warning of what is to
tial number of middle-incomecountries-the Bank come. Added to these factors are the civic calamity
will provide external finance mainly for mainte- in some countries and the harsh and prolonged
nance and restoration. Such finance will be linked economic adversity in many more. But where the
to policy and institutional reforms. road problem is now serious or is about to turn
* For countries with severe financial and institutional serious, the damage attributable to such factors could
constraints on maintenance (category III)-almost usually have been mitigated by more efficientinstitu-
exclusively low-income countries-the Bank's as- tional arrangements for the management of the coun-
sistance over the next few years will be devoted try's infrastructure. It is difficult to understand fully
with few exceptions to maintenance and restora- the causes of institutional failure in different coun-
tion. The assistance will be conditional on institu- tries-and even more difficult to prescribe counter-
tional reforms and on the allocation of road bud- measures guaranteed to cure the problem. But this is
gets exclusively to maintenance and restoration no case for leaving things as they are. Action is
(with only limited and defined exceptions). required, even if it has to be accepted as experimen-
* For countries with obsolete roads, technology, and tation. Bank support for the road sector of countries
organization (category IV), the Bank will empha- with serious road deterioration problems will there-
size organizational and technological improve- fore be guided by signs of institutional progress.
ments in modernizing the road networks and will These signs include the growth of political attention
also support new construction when it is justified. to the preservation of infrastructure, the strength-
In all this, the Bank will encourage the execution ening of internal accountability in the institutions
of maintenance work by entities outside the road charged with managing it, the introduction of incen-
authority-entities operating on commercial princi- tives that press people to perform, and the deliberate
ples, preferably in a competitive environment. In search for resources and schemes that promise care
addition, the Bank will promote measures to increase for hard-won national assets.
public awareness of the need for timely maintenance

4
 The State of the Roads

The developing world's road building boom in the road maintenance usually is not exposed to either
1960s and 1970s created an infrastructure that has the economic consequences of undermaintenance
been crumbling in the 1980s and threatens to collapse or the organized pressure for better roads and thus
in the 1990s if not quickly strengthened and protect- has little incentive to provide responsive, effective
ed. Large road networks, built at great expense, have maintenance.
been inadequately maintained and used more heavily The costs to road authorities are only the tip of the
than expected. The result in many developing coun- iceberg, for the costs to road users operating vehicles
tries is a network of deteriorating roads. Many roads on rough roads are much larger. High haulage costs
are in such poor condition that normal maintenance constrain the location of economic activity, hamper
is no longer sufficient or effective. These roads now the integration of economic markets, limit the gains
require rehabilitation or reconstruction at three to from specialization, and render unviable many activ-
five times the cost of timely preventive maintenance ities that rely on road transport (see box 1-1).
and strengthening. And many more roads, whose
deterioration is not yet visible, will soon reach that An Overview of Road Conditions
point if they are not better maintained.
The problems of poor maintenance are worse for This study assesses road deterioration in the eighty-
roads than for other sectors for three reasons: five developing countries receiving highway sector
* The costs and financial requirements are large. assistance from the World Bank-officially, the Inter-
For example, expenditures to compensate for past national Bank for Reconstruction and Development
omissions in preventive maintenance in Sub-Saha- (IBRD) and its affiliate, the International Development
ran Africa are at least ten times those of supplying Association (IDA). The analysis covers only the main
textbooks for all elementary schoolchildren in the road networks, both because they are the most im-
region until the year 2000. portant roads and because information on other
* Road deterioration accelerates with time. This roads is fragmentary. In these countries 70 to 80
phenomenon makes it difficult, but also all the percent of interurban traffic is on the main networks.
more important, to recognize the need for preven- These networks consist of 1.8 million kilometers of
tive maintenance before deterioration becomes ob- primary and secondary roads (of which 1.0 million
vious and more expensive rehabilitation or recon- kilometers are paved) and have a replacement value
struction becomes necessary. of about $300 billion, excluding the value of land,
* Road authorities are insulated from the effects of bridges, and major earthworks (see table A-1 in
undermaintenance. The agency responsible for appendix A). The main networks generally include

5
 Box 1-1. The Consequences of Road Neglect in Ghana

Road deterioration can make an economic crisis worse. timber industry. Logs moving from the Kumasi area for
In Ghana a good road network was built before 1970, export through the Takoradi Port are trucked over a 500-
but it later suffered from serious neglect. By 1984about kilometer route because the direct road, which is about
60 percent of the main paved roads were in a state of half as long, is broken down and the rail services are
moderate to severe deterioration. Important sections unreliable. The detour adds $15 to $20 for each ton of
have become almost impassable, and access to some logs exported.
interior areas has been severelycurtailed. Transporters The prospects are not encouraging. The resources
refuse to go there because they do nor want their vehicles needed for road resurfacingand strengthening(mainlyto
to break down. reseal 1,100kilometersof paved roads and regravel1,600
Transport costs have increased in real terms by about kilometers of unpaved roads) constitute only about 20
50 percent on main roads and by more than 100 percent percent of the sector's total annual requirements,which
on rural roads, which have sufferedevengreater neglecr. are estimated at $47 million. Even with a recent road
In many areas the market rate for transporting fertilizer maintenance project supported by the World Bank, the
is as high as a dollar per ton-mile. These high costs have allocation for maintenance will increaseonly to about 30
cut deeply into farm returns, particularly for poor farm- percentof what is neededby 1989.Unlessmore resources
ers in areas away from main roads: some villagers can can be put into road maintenance-and applied effec-
no longer move their cocoa stocks to regional depots. tively-deterioration will continue. If so, it will take far
And during the 1982-83faminepoor roads prevented the more costly rehabilitation to forestall the almost total
transport of food from surplus areas to areas facing collapseof road transport in Ghana.
starvation. The high cost of transport has also hit the

the principal roads and highways that cross urban no road should be allowed to decline into poor
areas or provide access to ports. The analysis does condition unless it is to be kept in that condition
not deal with urban roads; nor does it cover 5 million deliberately (with routine maintenance but no resur-
to 6 million kilometers of rural roads and tracks facing or rehabilitation) or abandoned entirely. Sec-
(most of which are unpaved and lightly traveled) ond, normally there is a period of about five to eight
with a replacement value of perhaps $75 billion to years during which paved roads in fair condition can
$100 billion (Faiz, Harral, and Johansen 1987). be restored by resurfacing or strengthening; after that
Detailed assessments of road conditions are sub- time more costly measures become necessary. The
ject to considerable error. The importance of system- existence of many roads in fair condition suggests,
atic monitoring and evaluation of road conditions therefore, that extensive maintenance is needed
has only recently become generally recognized, and quickly if these roads are to be saved before they
few developing countries have data bases and man- decline to poor condition. And third, the cost of
agement information systems that are adequate to operating vehicles (especially large trucks) rises as
the task. Nevertheless, the evidence is good enough roads deteriorate. Because vehicle operating costs are
to show the broad dimensions of the problem. Re- the largest part of transport costs on all but the most
cent field surveys, supplemented by the judgment of lightly traveled roads, the increase in operating costs
World Bank engineers, make it possible to distribute swamps all other costs at stake in road management
a country's roads among three classes of condi- as roads deteriorate.
tion: good, fair, and poor (Mason 1985; Mason Based on regional averages, road conditions in the
and Miquel 1986). A road in good condition requires countries studied are alarming on two counts (see
only routine maintenance to remain that way. A road table 1-1). First, more than a fourth of all paved
in fair condition needs resurfacing. A road in poor roads-some 269,000 kilometers-are already in
condition has deteriorated to the point that it re- poor condition and need rehabilitation or reconstruc-
quires either partial or full reconstruction (see the tion. Second, the heavy concentration of paved roads
glossary at the front of this book for definitions of in fair condition (42 percent) foreshadows a major
the various types of road maintenance and improve- crisis unless concerted efforts prevent these roads
ments). from deteriorating into poor condition. Less serious
Three facts about road deterioration help to clarify deterioration in the 1970s in the U.S. network of
the problem. First, because reconstruction costs three highways receiving federal aid prompted widespread
to five times as much as resurfacing or rehabilitation, alarm, new legislation, new user taxes, and a large

6
Table 1-1. Condition of Main Roads, by Region
(percent)

Paved Unpaved
Region Good Fair Poor Good Fair Poor

Eastern ancl Southern Africa 42 32 26 42 30 28

Western Africa 52 23 25 20 36 44
East Asia aind Pacific 20 59 21 41 34 25
South Asia 19 45 36 6 39 55
Europe, the Middle East, and North Africa 41 35 24 30 46 24
Latin America and the Caribbean 44 32 24 24 43 33
Average 32 42 26 31 36 33
United States (Federal Aid Network, 1981) 31 57 12 - - -
United Kingdom (Trunk Road System, 1983) 85 12 3

-- Not applicable.
Note: The percentages are weighted by the length of the main road networks in each country in the regions.
Sources: For the United States, U.S. Congressional Budget Office (1983), pp. 20-21; for the United Kingdom, U.K. National Development Office
(1985), p. 9; for the six regions, World Bank survey of eighty-five countries that was based, as far as possible, on published information about
pavement conditions in sixty of the countries and supplemented, where necessary, by the judgment of Bank highway engineers. See the glossary
for definitions.

infusion of federal and state resources in the 1980s Age is important to the condition of paved roads
(Baker 1984; U.S. Congressional Budget Office 1983). because of the time path of their deterioration. Typ-
The unpaved roads in the main networks are even ically, two-thirds of pavement deterioration (and an
worse. True, the timing of maintenance for unpaved even higher proportion of maintenance cost) is con-
roads is less critical than for paved roads because centrated in the final third of the design life of the
the costs of restoring unpaved roads are less sensi- pavement (see chapter 2). After a boom in road
tive to the timing of the intervention. Even so, construction, a grace period of several years-during
however, the aggregate effect of their deterioration which roads remain in good condition even without
on vehicle operating costs can be considerable. Al- maintenance-is followed by a period in which the
though unpaved roads normally carry much less need for maintenance surges dramatically.
traffic, the ride quality of an unmaintained unpaved Differences in the age of networks underlie the
road deteriorates many times faster than that of a regional differences shown in table 1-1. On the
paved road (except in the terminal phase of the whole, paved roads in Western Africa are in better
paved road's life cycle). Deferring routine mainte- condition than the roads elsewhere. The networks in
nance on unpaved roads can quickly double the Western Africa are fairly new: more than half the
vehicle operating costs. paved roads were constructed, upgraded, or recon-
structed in the past ten years.
Major Determinants of Road Conditions Nigeria and C6te d'lvoire have substantially re-
built their paved networks since 1975 (see box 1-2).
The considerable variation in road conditions from Other less well-enclowed countries in the region were
country to country and region to region stems from helped by external aid agencies to develop and im-
differences in the past maintenance needs of individ- prove their road networks. Of twenty countries in
ual networks and the countries' responses. Western Africa for which information is available,
fifteen have young networks. The proportion is not
Need for Maintenance much lower in Eastern or Southern Africa, where ten
of sixteen paved networks are fairly new. Without
The maintenance needs of a road network can be substantial external assistance, however, many of
predicted fairly accurately from a set of structural these networks are unlikely to remain serviceable
characteristics, such as age, climate, traffic, design beyond the next ten years.
standards, construction quality, and subsequent The growth of traffic on roads built to obsolete
maintenance. Of these, age, traffic, and construction standards helps to explain the general condition of
quality are of particular importance in the developing the roads in South Asia: only 19 percent of the paved
countries. roads remain in good condition, and 36 percent are

7
 Box 1-2. Nigeria's New Roads and the Risk of Massive Deterioration

Of the 21,000 kilometers of paved roads in the Nigerian precarious. Many of the roads were built with generous
federal trunk road system, 36 percent were constructed geometric features but weak pavements, which require
or rehabilitated in 1975-80 and another 24 percent in substantial strengthening.Without an extensiveprogram
1981-85.As a result 62 percent of the network is rated to resurfaceand strengthen about 2,000to 3,000 kilome-
good, 15 percent fair, and 23 percent poor. Neglectof the ters of pavement a year, at an estimated annual cost of
unpaved network has rendered an estimated 90 percent $150 million to $200 million, the Nigerian trunk road
of unpaved roads in poor condition. system may deteriorate rapidly and require massive
Despite the remarkable expansion and improvement rehabilitation and reconstruction in the next ten to
of the Nigerian trunk road systemin the past ten years- fifteen years.
at an estimated cost of $8 billion-the situation is still

in poor condition. South Asia's road networks con- Latin America, has seen serious erosion despite sub-
sist mainly of aging roads with neither the geometric stantial efforts. In 1984, 28 percent of the network
capacity nor the structural strength to carry current was in poor condition, up from 18 percent in 1979
traffic. Both the volume of road traffic and the axle (see box 1-3). The deterioration was even more
loads have increased over the past decade as econo- dramatic in Honduras, which had previously had
mies have grown and traffic has shifted from other adequate road maintenance. The percentage of the
modes. Many of the roads in these networks need paved roads in good condition there dropped from
upgrading in addition to normal maintenance and 82 percent in 1981 to 50 percent in 1984, largely as a
rehabilitation. Countries in South Asia, particularly result of financial constraints.
India, face the need for a very large program of road A broad indicator of the economic burden of road
building. maintenance is the ratio of the length of a country's
The age of networks and the growth of traffic road network to gross national product, with an
explain some differences in maintenance require- allowance for the lower traffic volumes in poorer
ments, but the present condition of the roads reflects countries. The ratio for main roads ranges from 0.3
the extent to which maintenance requirements have kilometer per million dollars of GNP for the Republic
been met in the past. Inadequate maintenance is of Korea and 0.4 for Nigeria to more than 8.0 for
largely the result of misallocated funds, unsound Botswana and Zaire and 14.5 for Guinea-Bissau. Of
maintenance strategies, and inefficient implementa- the thirty-six countries with the highest ratios, thirty-
tion. two are in Sub-Saharan Africa (see table A-2 in
appendix A). Even with the best management, these
Financial Capacity countries would still face the world's highest burden
of road maintenance requirements relative to income
The capacity to pay for road maintenance from and would be likely to have lower than average levels
domestic sources depends on a country's resources. of maintenance.
Gross national product (GNP) per capita may thus be Financing road maintenance by borrowing from
a good index of financial capacity, but some govern- commercial sources abroad or development institu-
ments are able to capture more of it than others. And tions could help to fill the funding gap, but many of
of the fiscal intake, the allocation to highways is a these sources have preferred to finance construction
political decision about national priorities. Varia- rather than maintenance.
tions in income growth also affect funding: severe
setbacks in income growth clearly explain some Institutional Capacity
important instances of underfunded maintenance
and consequent road deterioration. The foregoing indicators-the differences in GNP per
Road conditions in Latin America are worsening capita and its rate of change, the ratio of network
rapidly, mirroring the general economic downturn length to GNP, the age distribution of roads, and the
and accompanying financial stringencies of the past volume of traffic-do not fully account for the vari-
five years-and illustrating how quickly roads can ation in road conditions from country to country.
deteriorate once they reach the critical age. Brazil, Several higher-income developing countries have
the country with by far the largest road network in poor roads, and some of the lowest-income countries

8
 Box 1-3. A Case of Undermaintenance: Brazil's Federal Highway Network

A 1979 survey of Brazil's federal highway network The increase in the percentage of roads in good
showed the followingdistribution of road conditions: condition was the result of new construction, not good
Percentage maintenance:6,000 kilometers of new paved roads were
Kilometers of network constructed, while 2,000 kilometers of those formerly in
Good 10,000 24 good condition declined to fair. Nor did the massive
Fair 23,000 58 backlog of roads in fair condition receivethe resurfacing
Poor 7,000 18 or strengtheningthat was due. Some6,000 kilometers of
that group deteriorated to poor condition, which greatly
Restoration and preventivemaintenance of all roads at increased the number of kilometers and the percentage
that time would have cost $1.8 billilonin constant prices, of roads needing rehabilitation or reconstruction rather
A repeat survey in 1984rated the network as follows: thans peentivemaitance.
than justpreventivemaintenance.Tecostroftis
The costof thisresto-
resto-
Percentage ration is estimated at $1.7 billion, and the cost of the
Kilometers of network preventivemaintenance needed to save the roads still in
Good 14,000 30 fair condition is estimated at $750 million. Thus the
Fair 19,000 42 federal rehabilitation and maintenance backlog grew to
Poor 13,000 28 $2.4 billion, a one-third increasein six years.

have better ones. The capacity of a country to deal creased and systematic maintenance to prevent rapid
with its maintenance needs also depends on how deterioration. The road deterioration problem thus
effectively and efficiently it can use funds to protect pervades the developing world. At the core of the
and rehabilitate the road infrastructure-a factor problem are mainly the countries of Sub-Saharan
generally known as institutional capacity. Africa and South Asia, and at the hard core of the
Institutional capacity has several facets. One is the problem are those Sub-Saharan countries whose fi-
pool of skills, such as the size of the labor force, that nancial and institutional capacities are unequal to the
can be applied to the activity. Others are the sound- task at hand or to the one they will soon have to face.
ness of the maintenance strategy-type, level, and Although some networks in Sub-Saharan Africa
timing of intervention-and the managerial and op- (C6te d'Ivoire, Malawi, Niger, and Rwanda) are
erational efficiency with which the strategy is execut- among the better maintained, they are also of recent
ed. These depend, in turn, on such factors as vintage and will soon require much more mainte-
government commitment, institutional structure, nance as the paved roads pass into the critical stage.
managerial ability, staff quality, accountability, and In the rest of Sub-Saharan Africa, recently expanded
incentives. Experience with these aspects of institu- networks are also numerous and therefore in better
tional capacity has often been remarkably disap- condition than older ones. In ten countries with older
pointing. And where limited institutional capacity networks, an average of 44 percent of their paved
cannot readily be expanded, the prospects are poor road length is in poor condition: in the seventeen
for recovering infrastructure that has already dete- countries with fairly new paved roads, the propor-
riorated or, if the peak of maintenance needs is yet tion of poor roads is only 22 percent. There is,
to come, for preventing heavy losses of capital. however, no reason to believe that countries with
newer paved networks have better capacities or pol-
The Hard Core of the Problem icies for road maintenance than those with older
networks. If they did, their unpaved roads should be
Not all countries facing a maintenance crisis have in better condition; for without maintenance, un-
arrived there by the same trajectory. Some have paved roads deteriorate rapidly and at a uniform
backlogs of maintenance needs because their finan- rate. The condition of unpaved roads in countries
cial and institutional capacities have not expanded as with newer paved roads differs little from that in
fast as their road networks. Some have built up countries with older paved roads. So the state of a
backlogs by deferring needed maintenance during country's paved roads, by itself, says little about the
adverse economic conditions. Some do not appear to country's maintenance capacity.
have backlogs now, but their networks of relatively In many countries the rapid expansion of networks
recent construction will soon require greatly in- has outstripped the growth in institutional capacity

9
 Box 1-4. The Construction-Maintenance Tradeoff: Consequences for the Road Network

The tradeoff betweenconstruction and maintenance and If this construction-maintenancepolicy continues for
its implications for the road network can be explained ten years, 100 kilometers of new paved roads will have
by a simple construct. Considera main road network of been built, but lack of maintenance during this period
1,000kilometers of paved roads that requires an annual will have rendered400 kilometersunserviceable,because
budget of $6 million, or $6,000 per kilometer-$1.5 only 600 kilometers can be adequately maintained with
million for routine maintenance (at $1,500per kilometer the residual maintenance budget. The rehabilitation bill
per year) and $4.5 millionfor resurfacing(at $45,000per for these 400 kilometers at the end of ten years would
kilometer every ten years, or $4,500 per year). The amount to some $70 million, at an average cost of
average estimated remaining life of paved roads with $175,000per kilometer.Thus over a ten-yearperiod the
little or no maintenance is ten years. If this budget is road network will suffer a net dimunition of 300kilome-
partly used for new construction (at $240,000per kilo- ters: the gain of 100 kilometers of new roads as against
meter) to extend the road network by 1 percent each a loss of 400kilometers of existing roads through lack of
year, the following results will emerge: maintenance.
Annual increase in road length: 10 kilometers This analysis (adapted from Robinson 1987) assumes
Annual construction cost: $2.4 million that the funds availablefor maintenance and resurfacing
Balance of budget remaining for maintenance: are concentrated on a restricted length of 600kilometers.
$3.6 million If the funds were instead spread across the entire net-
Length of road that can be annually maintained work, the loss of serviceableroads would likely be even
and resurfaced with the residual budget, at greater.
$6,000 per kilometer: 600 kilometers.

for road maintenance. The full consequences of this paved roads are in good condition, whereas none of
lag are not yet evident in most paved networks, the unpaved roads are rated as good. In C6te d'Ivoire
because most of the roads are still in the grace period 78 percent of the paved roads, but only 30 percent of
during which they need little maintenance and show the unpaved roads, are classified as good. Both coun-
few signs of deterioration even when maintenance is tries developed their paved networks fairly recently,
neglected. Prolonged undermaintenance, however, so their maintenance requirements will grow signifi-
can eventually result in a net diminution of the road cantly in the next few years. Almost every region has
network (see box 1-4). some countries in or approaching crisis, but Sub-
The effects of undermaintenance are already re- Saharan Africa stands out as having the most coun-
flected in the relatively poorer condition of unpaved tries at the hard core of the road deterioration
roads in all regions. In Nigeria 62 percent of the problem.

10
 TechnicalOptions
and Their Economic
Consequences

A few fundamental relationships explain road dete- maintained, deteriorate rapidly and at a fairly uni-
rioration and its consequences for the total cost of form rate throughout their life cycle. But unmain-
road transport. They hold true in a wide variety of tained paved roads follow a distinctly nonlinear path.
environments and form the starting point of any (All references in this chapter to the absence of
rational maintenance plan. Recent research has maintenance assume that the minimal work of vege-
greatly clarified these relationships and has quanti- tation control and drainage clearing is carried out.
fied and refined them empirically. As a result, eco- These low-cost activities are vital to the continuing
nomical maintenance strategies can be designed for serviceability of roads, and neglecting them is a sure
a wide range of circumstances, and the error of way to hasten road deterioration.)
certain strategies can be seen. Above all, these rela- During a long initial phase that lasts up to two-
tionships and the new data make it possible to thirds of their life cycle, paved roads undergo little
provide firmer guidance for countries that have to visible deterioration. This is followed by a phase of
restore badly deteriorated parts of their networks increasing-and increasingly rapid-deterioration
while keeping the other parts serviceable. (into fair condition) that ends within a few years in
Economic decisions about a highway system must radical structural failure (poor condition). This non-
take into account the total cost of transport on the linear path of deterioration affects the choice of the
roads: the discounted life-cycle cost of constructing optimal maintenance policy for paved roads (see
and maintaining the roads and the (usually) far larger figure 2-1). For the unwary it also tends to disguise
cost of operating vehicles on these roads. These costs the future maintenance requirements of young net-
must be estimated on the basis of sound knowledge works.
about road and traffic conditions and the interactions During the first phase a paved road can be kept
between them, as well as the applicability of different in good condition with fairly inexpensive routine
maintenance techniques in different environments. maintenance. In the subsequent phase of increased
deterioration, the pavement can be restored to good
The Fundamental Relations condition by resurfacing or, at a moderate cost, by
adding an overlay. An overlay will restore the ride
As roads deteriorate they get rougher. Until recently quality of the road, make the pavement strong
roughness was assessed subjectively, but now it can enough to meet traffic requirements for the next
be quantitatively evaluated using measures such as several years, and thus start a new pavement cycle.
the International Roughness Index (IRI) (Sayers, Gil- With adequate routine maintenance and timely resur-
lespie, and Paterson 1986). Unpaved roads, if not facing or strengthening, a paved road should not

11
Figure 2-1. Deterioration of Paved Roads over Time

With axle loads

'~ 12 15 percent higher than normal With normal loading
E 11
10
Reconstruction or major rehabilitation required
;. 9 (in absence of earlier strengthening)

Critical stage in life

C 6 cycle of paved road
5

4 Pavement strengthening indicated

- 3

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Years

Note: Data are based on the following assumptions: asphalt concrete pavement (structural number =3.5); average daily traffic - 2,500 vehicles; rainfall =120

millimeters a month; minimal maintenance.

deteriorate into poor condition or require reconstruc- estimated. The new estimates are, on average, a
tion. quarter below those proposed in an earlier standard
Another fundamental relation links vehicle operat- work (see appendix B). Overestimates of the savings
ing costs to road roughness (see table 2-1). Recent in vehicle operating costs may have led to faulty
research has shown that roughness has less of an investment and maintenance decisions, but a far
effect on vehicle operating costs than previously graver error occurs when maintenance policy ignores
the share of vehicle operating costs in the total cost
of road transport. That share is high, except where
Table 2-1. The Effect of Road Roughness traffic is extremely light (see figure 2-2). Even a small
on Vehicle Operating Costs percentage increase in vehicle operating costs from
(Indexof vehicleoperatingcosts:good= 100,at 2.3 IRI) rougher roads will outweigh the cost savings from
deferring maintenance.
Road condition"

Vehicleclass Fair(4.61Ro) Poor(6.9-9.2iRi) Strategies for Road Investment and Maintenance

uSmlca 106 109-26 Much information about the process and effects of
Light diesel truck ill 124-38 road deterioration comes from empirical research
Heavy truck 114 129-46 supported by the World Bank (see box 2-1). The large
Articulated truck 112 127-44 data set represents an important contribution to the
effort to design cost-effective road strategies-espe-
Notes: Data are based on 1984economiccosts for Costa Rica. cially ones combining construction and maintenance
a. The dominant effect of road condition on vehicle operating costs
is through "roughness," a measure of road surface irregularities policies-for developing countries. A model based on
standardizedby the international RoughnessIndex (IRI). that research has been used in studies of road mainte-
Source: From application of the vehicle operating costs submodel nnei eea onre,icuigCie ot
of the HighwayDesignand MaintenanceStandardsModel (HDm-ill). nnei eea onre,icuigCie ot
For the Costa Rican context, see Bhandari and others (1987). Rica, and Mali, which as a group exemplify a wide

12
range of transport conditions found in the developing with satisfactory outcomes. The results, although
world (see appendix B). These studies demonstrate specific to the conditions in the countries studied,
the interdependence of investment and maintenance illustrate general points about the selection of cost-
decisions and the role of information on road condi- effective maintenance policies and the criteria for
tions, traffic, and traffic growth in ensuring decisions deciding whether or when to pave and what pave-
ment strength to choose. An annual discount rate of
12 percent is used for the analyses reported in this
Figure 2-2. Vehicle Operating Costs as a chapter (Bhandari and others 1987).
Percentage of Total Life-Cycle Transport Costs
for Typical Cases with Optimal Maintenance Trafficand Road Conditions
In establishing standards for road design and mainte-
nance, the key factor is the level of actual and
100 expected traffic. For Chile and Costa Rica, which
N w _~ have relatively high volumes of traffic on their road
75
D m~jf networks, the optimal policy was to keep the most
heavily used two-thirds of the networks in good
~- t50 condition while maintaining the other third at lower
standards. In contrast, the most economical solution
25
U XF for Mali, where traffic volumes are generally low,
was to keep about 1 percent of the network in good
044
A: l l l l l condition and maintain the rest at reduced standards
Q 0 1,000 2,000 3,000 4,000 5,000
as determined by traffic.
Initialaveragedailytraffic(numberof vehicles) Under these optimal policies, keeping roads in
less than good condition does not imply neglecting
Nore:Trafficis assumedlo grow at 3 percenta year.The presentvalueof road maintenance. In all cases drainage and vegetation
constructionand maintenancecosts + vehicleoperatingcostsis computedwith
a discountrateof 12percent. control are fundamental, as are high standards of
patching and basic routine maintenance, even for

Box 2-1. The Highway Design and Maintenance Standards Study

In 1971the WorldBankinitiatedwhat later becamea provementsto road geometry(Hoban1987).This re-
major collaborativeprogramof primarydata collection searchadvancedourunderstandingof the linksbetween
and research-firstin Kenyaand later,on a muchlarger road deteriorationand vehicleoperatingcosts beyond
scale,in Braziland India. More than $20 millionwas the state of ten or fifteenyears ago. It revealedthe
spentto collectandanalyzedata on highwayconditions complexityof the relations-which wereformulatedon
and vehicleoperatingcosts so that key relationships the basis of a more comprehensive data set than was
linkingroad construction,maintenance,and vehicleop- previouslyavailable-and theirsensitivityto an array of
eratingcostson differentroadsand underdifferentroad area-specific
circumstances.
conditionscouldbe quantified.Morethan 90percentof To apply this empiricalknowledgeto the related
the fundscamefromthe other participants:the govern- questionsof highwayconstructionand maintenance,the
mentsof Brazil,India,Kenya,and the UnitedKingdom, WorldBankdevelopedthe HighwayDesignandMainte-
as wellas the UnitedNationsDevelopment Programme. nanceStandardsModeland a companionExpenditure
Australiaand Swedenprovidedtechnicalassistance. BudgetingModel. These modelssearch for the best
Resultsof the studiesare reportedin a seriesof World solution appropriateto a country'scircumstancesby
Bank publications,including two volumeson road exploringthe effectof multiplecombinationsof road
user costs (Chesherand Harrison 1987;Watanatada, designand maintenance optionson total transportcosts.
Dhareshwar,and Lima1987),a volumeon road deteri- Theresultscanthenbe testedforsensitivityto variations
oration(Paterson1987),anotheron the planningmodel in input parametersand futureconditions-a necessary
(Watanatadaand others 1987),two technicalpaperson precaution.Successful applicationof eithermodel,how-
measuringroad roughness(Sayers,Gillespie,and Pater- ever,dependson the reliabilityof the informationsystem
son 1986;Sayers,Gillespie,and Queiroz1986),and a andthe abilityof staffto usethe data baseandthe model
technicalpaper on evaluatingtrafficcapacityand im- properly.

13
roads to be maintained at low standards. In Mali paved and unpaved roads. If, on the one hand, the
this maintenance is estimated to cost $6.2 million a availability of future maintenance funds is uncertain,
year-about twice the current spending. This esti- the best policy is to defer paving. The present value
mate, however, does not include the $9 million re- of the total life-cycletransport costs associated with
quired to rehabilitate and reconstruct the backlog of a gravel road (even if suboptimally maintained) will
higher-volume paved roads in poor condition that be less than that for a paved road if adequate funds
warrant saving. may not be available when the paved road begins to
deteriorate sharply. If, on the other hand, the coun-
To Pave or Not To Pave try's ability to plan and execute maintenance is in
doubt, early paving (and therefore fewer roads) is
Decisions about whether and when to pave a road indicated. The present value of the total life-cycle
must take many factors into account, including the transport costs associated with an unmaintained
expected growth in traffic. Figure 2-3 shows, for a gravel road will exceed that of an unmaintained
specific example in Mali, the net present value of paved road down to the point at which the roads
paving a gravel road as a function of traffic. Compar- cease to be serviceable. These conclusions, which
ison of cases 1 (no growth in traffic) and 2 (6 percent have been stated in the broadest terms, reflect the
annual growth) indicates that the traffic level at joint effect of discounting and the nonlinear path of
which it becomes optimal to pave the road is sensitive deterioration of paved (but not unpaved) roads (see
to the expected rate of traffic growth. It is assumed appendix B).
in all cases that proper maintenance will be carried
out whether the road is paved or not. The cost of Strong or Weak Pavements
paving is also important: in case 3 (no growth in
traffic but a 50 percent increase in cost), the optimal An important question in determining the paving
traffic threshold for paving jumps from 270 vehicles strategy is whether it makes economic sense to begin
a day to 425. But the penalty for not paving at the with a low-strength design and then strengthen the
optimal traffic level is not large, and considerations pavement later to accommodate heavier traffic and
other than traffic may govern the choice. axle loadings. The answer depends on the quality of
One consideration is whether future maintenance maintenance that can be expected on the (initially)
is likely be done to acceptable standards, on both lower-strength road. The question was explored for
conditions in Costa Rica and Mali. The studies,
which assumed a 3 to 4 percent annual growth in
Figure 2-3. Break-evenTrafficVolumes for traffic and two levels of axle loads, indicated that
Paving a Gravel Road with a 75 to 90 percent probability of adequate
maintenance, staged construction would be econom-
. 100- ical for initial traffic flows of up to 2,000 or 2,500
Ej ^ Case 2,' vehicles a day, depending on axle loads. But with
, g ,-' ~~~~~~~~~~only
a 30 percent probability of adequate mainte-
nance, strong pavements should be built initially,
E 50 despite the higher construction cost, to compensate
- m Case 1 for unreliable maintenance; in this case staged con-
G - , Ca_ struction would be the preferred alternative only at
155
n 4 ,/ 270 - ~ traffic volumes of less than 1,000 vehicles a day with
>cc 0 _42 light axle loads. In all cases the pavement should
X ..- .- " __425 conform to minimum design and construction speci-
,, .- Case fications. A technical option that has not been suf-
G~ o _ Case 3. ficiently explored in developing countries, particu-
Z -50 I I I I larly in the humid tropics, is the use of pavements
0 100 200 300 400 500 made of portland cement. Although they are costly
Initial average daily traffic (number of vehicles) to construct, concrete pavements can provide an
initial service period of fifteen to twenty years that is
Note: Case I (the base case) assumes no growth in traffic and optimal nearly maintenance free.
maintenance policies on paved and unpaved roads; case 2 assumes a traffic
growth rate of 6 percent a year and the same maintenance policies as in case 1; Some countries made expensive mistakes by stag-
case 3 assumes no growth in traffic and a 50 percent increase in paving costs. ing the construction of pavements under assumptions
of adequate maintenance and restricted axle loads

14
that proved invalid. Heavy loads accelerate road axle loads, because the economic gain from the use
deterioration, and even with optimal maintenance of larger trucks outweighs the increased damage to
the average life-cycleroughness remains higher with roads (Faiz and Fossberg 1987;Rolt 1981).
heavier axle loads. Empirical evidence shows that
pavement damage increases exponentially (to the Tactics under Budgetary Constraints
power of 4) with increased axle loads. But the regu-
lation of axle loads has proved exceedingly difficult Road authorities frequently do not have enough
and expensive.Many road authorities are now build- funds for the economically optimal level of mainte-
ing stronger, more expensive pavements than would nance. When budgets are constrained, the best policy
be necessaryif axle loads were effectivelycontrolled. is not simply to reduce all categories of maintenance
Even if load regulations could be enforced, the limits spending equally, as is often done. The situation calls
should be increased beyond the prevailing eight- to for revising policies and using different maintenance
ten-ton single (and thirteen- to sixteen-ton tandem) options. The management of retrenchment (and, in

Figure 2-4. Net Present Value versus Road Agency Expenditures for Alternative Maintenance Strategies
for a Paved Road in Fair Condition
(thousandsof dollarsper kilometer)

180
40-millimeter
overlaysat 3.5IRI (H) 80-millimeter
overlay(I)
160 _ 40-millimeter
overlaysat 4.2IRI (G)

0- illimeter overlaysat 5.0IRI(F)

140 -

; 120 - Resealing(D)
Patchingonly * Reconstruction
only (nopatching)(E)
-E (B)
. 100 * Patchingandreconstruction(C)

80

60
zX

40

20-

No pavementmaintenance(A)
0 10 20 30 40 50
Presentvalueof road agencyexpenditures

Note: Averagedaily traffic = 800vehicles.SeeappendixB for specifications of themaintenancealternativesrepresentedby each point on the graph. It is possibleto
operate on the frontierbetweentwo of thepoints byusingone policyon someroads and theother policyon others.For example,a combinationof policiesF and G
coulduse fundsequal to thoserequiredfor policyE and achievea net presentvalueon the frontier.Thus policyE is inferiorand shouldneverbe used.Similarly,any
other points belowthe efficiencyfrontierrepresentpoliciesthat shouldnot be used.In the caseillustrated,policyD (resealing)is inferiorto a strategycombiningthe
B (patchingonly)and F (overlay)options.In other cases,resealingwill often turn out to be an efficientway to prolongthe functionallife of pavements.The relative
positionof variouspolicieswill changesignificantlywithchangesin the discountrate and the relativecost of differentmaintenancealternatives.

15
extreme cases, network contraction) has not yet been agency and somewhat higher costs to road users. In
thoroughly analyzed, but some measures have been general, such tradeoffs were found in the vicinity of
found to be effectivein controlling the retrenchment the optimal strategy in most of the other cases
process and preventing it from becoming haphazard. studied.
Typically, when spendingpriorities are established But as the budget departs further from the opti-
the tradeoffs between road agency costs and the net mum and maintenance is cut back correspondingly,
benefits of different maintenance strategies are not vehicle operating costs increase by much more than
examined. As long as the net present value of a the savings to the road agency. In addition, the
maintenance option is positive, it is deemed justifi- changes in maintenance policy entail increasingly
able-whatever the cost to the agency. But if the greater risks of pavement failure, which means much
analysis is extended to road agency expenditures, higher costs for reconstruction in the future. In sum-
there appears a big margin for reducing them without mary, the range of good maintenance policies is wide,
making large cuts in the benefitsto road users. Figure but below a certain level of funding the consequences
2-4 shows the net present value of alternative mainte- of further reduction become destructive.
nance options applied to a specificclass of roads. The Sometimes a government must take austere mea-
line connecting the points that are highest in net sures to deal with a national emergency or adjust to
present value for given agency expenditures is the economic conditions. It may cut back budgets for a
efficiencyfrontier. The optimal strategy, if unlimited year or several years, with the intention of restoring
funds and a discount rate of 12 percent are assumed, them to normal levels later. In such circumstances
would be that shown at point I. In addition to basic road maintenance is often deferred. If roads are in
routine maintenance, this strategy requires the imme- good condition, the deferral may not be too costly.
diate application of an 80-millimeter overlay, to be Increased roughness caused by undermaintenance
followed by 40-millimeter overlays whenever rough- will increase vehicle operating costs somewhat, but
ness exceeds 3.5 IRI. roads that decline to fair condition can be restored
The efficiencyfrontier is rather flat immediately to with a modest additional expenditure. If the deferral
the left of the optimum point; alternatives in this continues too long, however, or if roads are in only
zone can significantlyreduce agency costs with only fair condition to begin with, the impact on vehicle
a small impact on total transport costs. As agency operating costs will be greater. And paved roads that
expenditures are cut back, vehicle operating costs decline to poor condition will require far more ex-
increase, but by only a little more than the savings in pensive reconstruction later on. That is why road
maintenance expenditures. Moreover, the tradeoffs agenciesshould take deliberate account of the risk of
in the vicinity of the optimum do not imply higher future budget interruptions when making choices
future costs for road rehabilitation. It is simply a about road design and maintenance. Good mainte-
matter of maintaining the roads at a somewhat nance in normal times is one way to minimize the
rougher level, with a consequent saving to the road impact of future interruptions in funding.

16
 The Institutional Challenge

Choosing the best maintenance policy will remain an Structureand Functionsof the Road Agency
academic exerciseuntil institutionscan efficientlyput
the policy into praLctice.Past experienceis not encour- World Bank efforts to improve the management and
aging. The World Bank advanced more than $1.2 maintenance of roads have focused on road authori-
billion between 1971 and 1985 for training, technical ties, which typically operate as public sector monop-
assistance,and managementconsultanciesto improve olies. These agencies are generally responsible for
the organizationand managementof road administra- three functions: (a) planning the development and
tion in developingcountries. Other lendersand donors maintenance of roads, (b) negotiating with, and over-
and the developingcountries have also devoted sub- seeing the work of, contractors, and (c) constructing
stantial resources to this purpose. Despite these ef- and maintaining roads using their own work forces
forts to improve operational and administrative (force account). The division of work between con-
performance, it has been difficult to establish self- tractors and force account varies by country. For the
sustaining institutions that can manage road mainte- most part, however, contractors undertake new con-
nance efficientlyor use external resourceseffectively. struction and large rehabilitation projects, and the
Experience has provided no standard solutions to road agenciestake care of routine maintenance them-
the problems of institutional performance. Without selves. Resurfacingoperations are sometimes shared,
proven formulas, institutional development has had sometimesnot. In some countries contractors also do
to proceed by continuous, local experimentation. routine maintenance, and in others the agency's work
Experience has, however, identified constraints on force also constructs new roads.
improved performance and some general principles Contracted works for new construction and reha-
worth pursuing (Harral 1987). bilitation normally constitute the greatest part of
an agency's expenditures. But direct execution of
The Constraints maintenance accounts for the largest share of road
authority employment and for a disproportionate
Three factors have worked against the development share of difficulties.
of effectiveinstitutions for road maintenance: Conflicts may arise when a state organization
* The nature and constitution of the typical road combines the planning and control of construction
agency. Most road authorities have conflicting and maintenance operations with the direct execu-
objectives and functions and therefore operate tion of works. Internal control is blunted and efficien-
under incompatible incentives. cy prejudiced when direct execution receives too
* The weak public pressure for better roads. much emphasis, if only because of the political sig-
* The inadequacy and unreliability of funding. nificance of the large labor force employed. Already

17
weak incentives for efficiencyare undermined by the the budget goestoward the wages of an unnecessarily
constraints of public service. Civil service rules cir- large labor force, the disposable portion fluctuates
cumscribe managerial decisions. Deficient salary much more than the total when budgets are cut or
scales and promotion systems make it difficult to prices rise. A small cut can bring much of the agen-
retain competent staff. Employment objectives fre- cy's work to a standstill if, for example, it means the
quently smother the technical work of the authority agency cannot buy fuel or spare parts. Moreover,
and distort its decisions (box 3-1). even when funds have been appropriated for mainte-
nance, political or private interests put pressure on
Weak PublicPressure road authorities (or their financial sources) to divert
funds to other purposes. Such interference, often one
Because road agencies do not operate transport ser- of management's biggest challenges, can be con-
vices, they do not suffer directly from bad mainte- trolled only in the political arena.
nance. Nor do they normally sell services to road
users in a competitive market. They thus do not bear The Search for Solutions
the full cost of neglected maintenance; nor are they
subject to market pressure, as railways are. Also Efforts to improve the efficiencyof institutions for
absent is the pressure from direct dealings with the road management must focus on ways to make these
public-pressure felt by public health, education, institutions accountable for their performance.
and transport services, for example. Truckers rarely
experience enough competition from other transport InternalAccountability
modes to complain about the high cost of operating
vehicleson bad roads, and car owners and businesses The absence of internal accountability is a common
are too dispersed to form effective pressure groups cause of the institutional failure of road agencies.
for better roads. Furthermore, the effects of neglect- Planning, supervision, and execution are normally
ing maintenance are unlikely to be perceived before combined in the same organization. Road agencies
the problem has become acute. Without pressure seldom link the allocation of funds to explicit physi-
groups to prod them, politicians and administrators cal plans and rarely perform postproject evaluations.
have little concern for the roads, and the road agency To improve accountability, planning and supervision
becomes a facility for political patronage and unem- must be separated from the execution of works, and
ployment relief. procedures must be introduced to ensure that the
projects are evaluated. One way of doing this is for
InadequateBudgets the road agency to award maintenance contracts
through competitive bidding and then to supervise
Lacking political support and shielded from market and evaluate the work closely. Raising the visibility
pressures,the budgets for road maintenance are often of maintenance objectives and performance is equal-
unrelated to known requirements. Because much of ly important. Unless this is achieved, the chronic

Box 3-1. Overstaffingand Resource Imbalancesin Kenya

The effectsof the generalbudgetarysqueezeon operat- Three yearslater,however,the situationhad deterio-
ing funds(for fuel,spareparts, and bitumen,for exam- rated dramaticallybecauseof increasesin the sizeof the
ple) have been exacerbatedin Kenyaby two parallel permanentwork forceand decentralization of the road
developmentsaffectingthe Roads Department.One is agency's functions.The Roads Department'scasual
the decisionto integratecasuallaborwith the permanent workers(and somepersonnelfrom other departments)
work force; the other is the implementationof the had beenintegratedwith the permanentwork forceat
DistrictFocusProgram,a broadlybasedeffortto decen- the districtlevel,andthe numberof personnelhad risen
tralize governmentfunctions.Until 1982the (central) to 14,600.Morethan 90 percentof the K Sh210million
Roads Departmentemployedan averageof 1,000per- allocatedfromthe centralbudgetto thedistrictswentfor
manentstaffandabout 9,500casualworkers.Wagesand personnel-relateditems, which left little for operating
emolumentsamountedto 134millionKenyanshillings expenses.Meanwhile,funds for operationsthat were
(K Sh), and operatingfunds totaled K Sh194million, centrally controlled (includingmost equipment-based
whichwas alreadyinadequate. operations)weredownto K Sh126million.

18
underfunding of maintenance in developingcountries reform. Decentralization is seen as a way to make the
will be difficult to reverse. road authority more responsiveto local needs and to
Also essential for an agency's accountability is reduce the difficultiesof managing activitiesthat are
monitoring by an independentauthority. Performance geographically dispersed. An argument against de-
audits should relate financial flows and physical per- centralization is the fear that technical and manage-
formance indicators to the state of the roads. A system rial staff would be spread thin or that maintenance
for collectingthe data needed to monitor performance funds would be diverted to more politically popular
is therefore equally necessary. The data can also be (or more conspicuous) but less worthwhile construc-
used to make an independent and public assessment tion activities. Decentralization could also lead to the
of the agency's performance. In the United Kingdom, employment of excessivelocal staff-as in Kenyaand
for example, county road authorities have to publish Honduras. The Bank's experience is not conclusive
annual reports that show their costs in comparison on these points and finds little correlation between
with private contractors' charges (Cox 1987). success in road maintenance and the degree of cen-
tralization. Thus case-by-caseexperimentation seems
PublicAwareness the only practical prescription.
More obvious is the need to separate the execution
Greater public awareness, including that of potential of works from the planning and control functions, for
interest groups (contractors, exporters, transport en- this would help to insulate planning and control from
terprises), is important for shaping policy and mobi- the pressures of operating a works program. If exe-
lizing support for programs to restore and maintain cution were turned over to private contractors, the
roads. The Swedish National Road Administration, road agencies would be freer to concentrate on pro-
for example, conducts periodic surveys (using ques- curing and managing resources. Private contractors
tionnaires)to find out what the public thinks about its operating under the incentives of a competitive envi-
work. It uses the results to redirect its policies and ronment offer better prospects for developing an
operations and to persuade Parliament and the gov- efficient, lasting institutional capacity for carrying
ernment to keep highway funding at adequate levels. out road maintenance services. They can also consti-
The public information programs of professional so- tute an informed group with a direct interest in the
cieties and trade associations-such as the American adequacy of road maintenance budgets. Moreover,
Societyof Civil Engineers,the American Automobile experience suggests that private competitive organi-
Association, and the National Asphalt Association- zations use and maintain machinery more effectively
have helped to arouse public concern about the state than do government agencies (box 3-2) and can
of roads in the United States (Choate 1983).Perform- perform work at lower cost.
ing a similar role in France,the national and regional A review of experience with the contracting of
associationsof contractors (FederationNationale des maintenance in nine countries found that roads un-
Travaux Publics and Federation Regionales des Tra- der contract were generally well maintained in sev-
vaux Publics) lobby for increased public spending on en of them-Argentina, Brazil, Central African Re-
roads. In Japan, the Road Association,an interprofes- public, Ghana, Kenya, the United Kingdom, and
sional body with private and public representation, Yugoslavia (Harral, Henriod, and Graziano 1986).
promotes public interest, both technical and general, The problems encountered in initial experiments in
in the condition of the country's roads and acts as an the other two (Colombia and Nigeria) are being
effectivelobby at all levels of government. evaluated. By and large, contract maintenance tends
to be more cost-effectivethan maintenance by direct
Elements of Reform labor, and contractors have been attracted to mainte-
nance opportunities., even in remote areas. In many
Management systems that work in developed coun- cases small local contractors can operate in such
tries have not always improved the operation of road areas more cheaply than the central road administra-
maintenance institutions in developingcountries. Re- tion or larger contractors.
forms will thus have to be experimental (Harral 1987; The desirability of contracting a major part of
Faiz and Bennathan 1988). routine maintenance and resurfacing operations de-
serves wide consideration. A small government
Decentralizationor FunctionalSeparation maintenance capacity can be retained to ease the
transition, provide backup for emergencies,and re-
Decentralizing the organizations for road mainte- duce the risk of replacing a public monopoly with a
nance has often been advocated as a first step in cartel of private interests (as occurred in Nigeria and

19
 Box 3-2. Underutilization of Equipment in Western Africa and Latin America

Road maintenance surveys by the World Bank in 1985 countries in 1982were only about one-third the amount
found that vehicles and equipment were seriously needed for 1,000 hours of operation. The lack of spare
underutilized in Western Africa and Latin America. parts hindered maintenance, and the shortage of fuel
Utilization rates in these two regions were far below the restricted operation.
1,250 hours of operation a year that is generally regard- Low utilization is harder to explain in Latin Amer-
ed as efficient. In Western Africa the average annual ica. With three exceptions, the road authorities sur-
hours of operation ranged from 420 for rollers to 840 veyed were able to purchase enough spare parts to
for dump trucks. In Latin America utilization averaged operate the fleets 1,000 hours a year. About 40 percent
750 hours for equipment and 800 for vehicles; crushing of the agencies had less than two-thirds of the funds
plants were utilized even less, and asphalt finishers had needed for fuel for 1,000 hours of operation a year. But
the lowest rate-420 hours a year. the rest. did sufficiently better to raise the regional
In Western Africa the primary cause seems to have average to 82 percent of the funds needed for fuel. Poor
been a shortage of spare parts and fuel. About 40 equipment management and, possibly, a shortage of
percent of the road authorities received only half the qualified operators and mechanics may partly explain
funds needed to purchase enough spare parts to keep the low rates of utilization.
their fleets fully operational. Fuel expenditures in nine

to a lesser extent in Brazil). Japan's policy of relying If well-established domestic contractors are not
on contract maintenance is reinforced by numerical available, the trial introduction of small contracts
limits on the labor force that government agencies (say, for routine maintenance that is technically sim-
may hire. An increase in the demand for civil works ple and requires little investment) can reduce risks
leads automatically to the use of contractors. Even and help develop the capabilities of government
the field offices of the local public works administra- and contract personnel. All successful contracting
tions can let maintenance contracts of a limited, but schemes have involved close coordination between
far from negligible, value. Publicly owned equipment the government and contractors in defining and plan-
is kept at minimal levels, just enough to meet emer- ning the work. Such schemes have also changed the
gencies, so that the investment in government-owned role, sources, and contractual modes of foreign tech-
workshops and inventories of spare parts is very nical assistance.
small. The United Kingdom has had promising initial
results with a program (begun in 1981) that requires Human Resources
the agency's work force to compete with private
contractors (Cox 1987). In this way both sides are Better use can be made of human resources in several
encouraged to be efficient and to assemble reliable facets of road maintenance. Unskilled labor can be
cost data. substituted for machinery-a cost-effective solution
when wages are low, activities are properly orga-
Contracting Techniques and Skills nized and managed, and suitable incentives are built
into the system (Coukis and others 1983). And many
The use of contractors can reduce the burden on the countries have had good results with the simple
road authority, but it increases the need for supervi- "lengthman" system. Under this system people living
sory management. Harnessing the profit motive to alongside a road are responsible for maintaining it,
get the work done well, on time, and at reasonable and their payment and continuing employment are
cost requires knowledge of contract design. The contingent on satisfactory performance.
contract should not provide uneconomic incentives Low and inflexible civil service salary scales make
(as in a cost-plus contract) or restrict the contractor's it difficult for government agencies to retain compe-
freedom to seek cost-saving methods and sources of tent managers, engineers, technicians, foremen, me-
supply. Where institutional capability is limited, out- chanics, and others with special skills. Sometimes
side management consultants and specialists can help incentives can help an agency hold on to its most
to develop management systems and contract instru- productive staff. The road authority in Ethiopia,
ments-as well as provide training for government after some trial and error, devised a bonus scheme
staff and contractors. based on work-unit productivity. Agencies in several

20
other countries use special allowances to encourage agement systems, and training for road authority
field supervision by middle-level managers. But de- personnel. Technical assistance personnel usually act
spite these and similar schemes, staff turnover is as advisers and rarely take part in line operations. As
likely to continue to be high. Road authorities may a result, providers of technical services have little
as well recognize their usefulness as a training and exposure to risk, few incentives for improving per-
proving ground for personnel who will move on to formance, and only indirect responsibility (if any) for
the private sector. measurable output. Among the exceptions are the
Personnel management-or mismanagement-is technical assistance programs of Australia, France,
the area with the most potential for improvement. and the United Kingdom, which have provided tech-
Staff must be motivated and inculcated with a sense nical staff for senior line positions in the highway
of duty and accountability; good performance must agenciesof such countries as C6te d'Ivoire, Djibouti,
be recognized and rewarded. Recruitment and ap- The Gambia, Malawi, Niger, and Papua New Gui-
pointment to senior positions should be based on nea. The effectivenessof such assistance is greatly
competence, not patronage, seniority, or other anti- enhanced if the technical ministries of the sponsoring
quated civil service regulations and policies. Weak countries provide adequate backstopping.
performance stemrsnot from inadequate training- Some schemeshave tried to establish performance
on which it is all too often blamed-but rather from incentives for suppliers of technical assistance by
inadequate, nonexistent, or unenforced personnel redefining their role and leaving them more of the
management policies. Although its importance can- risk. But the traditional suppliers cannot generally
not be emphasized enough, training must be an afford to take significant risks. So it may be desirable
integral part of personnel management. Since the to look for new sources of assistance, such as inter-
work of road maintenance is dispersed throughout a national civil engineering contractors. These com-
country, an effectiveand fairly low-cost way to train panies are accustomed to risk-taking ventures and
maintenance staff is through road extension services performance-related incentives. Their staff have the
modeled on the training and visit system used in requisite qualifications and are used to working as
agriculture (Benor, Harrison, and Baxter 1984; management teams in developing countries.
Transportation Research Board 1986). Another option is a "twinning" arrangement be-
tween a developing country's road agency and a
Technical Assistance partner institution in a more developed country.
Such arrangements can facilitate regular exchanges
Technical assistance for road maintenance has tradi- of middle-level managers and technical staff. They
tionally been concentrated on advisory services,man- can also provide for the transfer of technology,

Box 3-3. Twinning of Turkish and U.S. Highway Organizations

In 1947the UnitedStatesinitiateda programto help When the technicalcooperationprogram began in

Turkeydevelopa highwayorganizationcapableof con- Turkey,local counterpartscouldnot be foundfor many
structingand maintainingan expandingroad system of the positionsin the technicaldivisions,suchas Ma-
(U.S.Departmentof Transportation1976).The U.S. terialsor Surveyand Design.To fill this gap, training
PublicRoads Administration(PRA)was entrustedwith programswereorganized,and the bettergraduateswent
thistask. Duringthe firstphaseof the programthe PRA on to becomeinstructorsfor new classes.The Turkish
madean assessmentof highwaydevelopment in Turkey personnelgainedpracticalexperienceas they built and
and concludeda formal agreementwith the Turkish maintainedthe highwaysystem,and when recruiting
governmenton the objectivesof the program and ar- satisfactoryequipmentoperators proved difficult,the
rangementsfor its implementation. TurkishArmysuppliedmenandofficersfortheprogram.
PRA staffinitiallyprovidedthe core organizationfor The last U.S.adviserleft Turkeyin 1958.Bythen the
the work.Specialistswereaddedas required,but as soon TurkishGeneralDirectorateof Highwayshad evolved
as Turkishpersonnelcould take over an organizational into a first-classorganizationcapable of maintaining
unit,that unit wasdroppedfromthe technicalassistance 27,000milesof all-weathernationalandprovincialhigh-
program.The firstdivisionto be transferredto exclusive ways, most of which were improvedor constructed
Turkishadministrationwas the Planningand Program- duringthe periodof U.Stechnicalcooperation.
ming Division.

21
procedures, and information systems (Cooper 1984). These permit the transfer of appropriate technology
The Federal Highway Administration of the United and cost-effective management systems and allow
States has collaborated in such arrangements with countries to share their experiences with institutional
the highway authorities of Argentina, Ethiopia, Jor- reform. The road maintenance conferences orga-
dan, Liberia, Nepal, Philippines, and Turkey, among nized by the regional U.N. agencies (the Economic
others (see box 3-3) (U.S. Department of Transporta- Commission for Africa, the Economic Commission
tion 1976). France, through its technical services for Latin America and the Caribbean, and the Eco-
agencies, has provided similar assistance to road nomic and Social Commission for Asia and the
institutions in Algeria, Cameroon, and Cote d'Ivoire. Pacific) have played an important role in fostering the
Equally important are technical cooperation and exchange of road maintenance experience among
exchange programs between developing countries. developing countries.

22
 FinancialRequirements

Few developing countries have spent wisely, or $15 billion to $25 billion. Under the constraints of
enough, on road maintenance. Not all roads have to limited budgets, only some of these components
be kept in the best:condition to maximize economic would qualify as economically warranted. The esti-
returns, but heavily used roads should not be al- mate also excludes about $3 billion for partial reha-
lowed to deteriorate-especially when a fairly bilitation of low-volume main roads in the poorest
small, properly timed expenditure for maintenance countries of Sub-Saharan Africa. Full restoration of
would make reconstruction unnecessary. The costs these roads would not be economically warranted,
for restoring deteriorated and rapidly deteriorating but they should be maintained to higher standards
roads are substantially more than most developing than they are now.
countries have been spending. But far greater sums
are at stake if the situation is not handled better over Future Costs
the next decade.
Strongly affecting future costs will be the money
The Overall Picture spent (and effectivelyused) in the next few years to
prevent more roads from deteriorating into poor
The financial requirements of the road deterioration condition. A high proportion of the paved network
problem have two parts. One is the cost to restore (more than 40 percent in 1984) is or will soon be in
(rehabilitate or reconstruct) those roads that are in only fair condition and will therefore require exten-
poor condition and warrant saving. The other is the sive maintenance to prevent structural failure. It is
future annual cost of maintaining the whole network estimated that resurfacing and routine maintenance
at economically warranted standards (Smith and needs will amount to about $4.7 billion a year (table
Harral 1987). 4-2) over the next few years. If these needs are not
met in time, however, the costs will multiply rapidly.
The Cost of Restoration An example of inadequate spending is the $43 billion
worth of restoration needed now because some $12
The total cost to restore the backlog of degraded billion of preventive maintenance was neglected over
main roads that *warrantsaving in the eighty-five the course of a decade or so.
developing countries is about $43 billion (table 4-1).
Even without a claim to accuracy, this number pro- Required Allocations
vides an order of magnitude. Bridges and large ter-
tiary and lower-order networks are not included, but The funds needed to prevent further deterioration
a speculative estimate for such components is about and to clear the $43 billion backlog of rehabilitation

23
Table 4-1. Total Expenditures Required to Meet the Estimated Restoration Backlog in 1984
(billions of dollars in 1986 prices)

Rehabilitation
and reconstruction Total as a
Paved Unpaved percentage of GNP
Region roads roads Total Median Range

Eastern and Southern Africa 1.4 0.8 2.2 3.3 0.0-15.5

Western Africa 1.9 0.9 2.8 3.2 1.2-36.9
East Asia and Pacific 7.5 1.8 9.3 1.6 0.2-2.5
South Asia 7.7 0.9 8.6 3.5 2.5-12.7
Europe, the Middle East, and North Africa 8.4 0.9 9.3 2.6 0.4-7.4
Latin America and the Caribbean 7.9 3.1 11.0 2.0 0.4-22.8
Total 34.8 8.4 43.2 -
Percentage of total 81 19 100

- Not applicable.
Note: The estimation procedures are described by Smith and Harral (1987). A few countries in each region account for a large proportion of
the region's restoration needs-Eastern Africa: Madagascar + Zambia = 40 percent; Western Africa: Nigeria + Ghana = 58 percent; East Asia and
Pacific: China = 56 percent; South Asia: India = 64 percent; Europe, the Middle East, and North Africa: Turkey + Yugoslavia + Algeria = 56 percent;
Latin America and the Caribbean: Brazil + Argentina = 63 percent.

Table 4-2. Annual Maintenance Expenditures Required to Prevent Deterioration, 1986-90

(billions of dollars in 1986 prices)

Resurfacing Total as a
Routine Paved Unpaved percentage of GNP
Region maintenance roads roads Total Median Range

Eastern and Southern Africa 0.2 0.1 0.1 0.4 0.7 0.2-1.8
Western Africa 0.1 0.1 0.1 0.3 1.0 0.1-3.5
East Asia and Pacific 0.3 0.8 0.2 1.3 0.2 0.1-0.4
South Asia 0.1 0.3 0.1 0.5 0.3 0.2-1.1
Europe, the Middle East,
and North Africa 0.2 0.5 0.1 0.8 0.3 0.1-0.6
Latin America and the Caribbean 0.5 0.6 0.3 1.4 0.4 0.1-2.8
Total 1.4 2.4 0.9 4.7 -

Percentage of total 30 51 19 100

-Not applicable.

needs (in either five or ten years) are summarized in poor condition would be greater because of the
table 4-3. If all countries were to meet the five-year savings in vehicle operating costs.
target, the cost, reckoned without discounting, would Costs could be reduced by improving the produc-
be about $13 billion a year ($4.8 billion for mainte- tivity of maintenance, but it is not clear whether such
nance and $8.6 billion for restoration). For the ten- gains are attainable in the short run. Although there
year target the cost would be about $9 billion a year is scope for improvement in the areas of labor,
($4.8 billion for maintenance and $4.3 billion for equipment, and procurement and supply, such gains
restoration). The foreign exchange requirements are likely to be disrupted if the institutional structure
would be $6 billion for the five-year target and $4 is defective. Even when supported by an infusion of
billion for the ten-year target. technical assistance, such gains are difficult to sustain
These estimates are less than economically optimal without the commitment and dedication of local staff
because they emphasize holding down road agency and managers of road authorities. Productivity gains
costs at the expense of higher costs to the users are likely, therefore, to account for only a small
(especially in the ten-year program). This emphasis reduction in the short-term bill.
gives priority to saving roads in fair condition to Total road expenditures, including those for new
minimize future restoration costs, even if the social construction, in the eighty-five countries are estimat-
return from reconstructing heavily traveled roads in ed at $10 billion to $13 billion for 1984, and restora-

24
Table 4-3. Annual Financing Requirements for Restoration and Maintenance of Roads
(billions of dollars in 1986 prices)

Total cost Foreignexchange

Ten-year Five-year Ten-year Five-year
Region target target target target

Eastern and Southern Africa 0.6 0.8 0.3 0.4

Western Africa 0.6 0.9 0.3 0.5
East Asia and Pacific 2.2 3.1 0.9 1.3
South Asia 1.4 2.3 0.6 0.9
Europe, the Middle East, and North Africa 1.8 2.7 0.8 1.3
Latin America and the Caribbean 2.5 3.6 1.2 1.7
Total 9.1 13.4 4.1 6.1

tion and maintenance may have accounted for about ing to their adequacy for restoration of the road
50 percent of this. With $4.8 billion a year required network:
for future maintenance, little would be left for catch- Sufficient financing capacity. Current road fund-
ing up on the rehabilitation backlog. In many coun- ing, if reallocated, would be sufficient for adequate
tries the restoration and maintenance budget is too maintenance of the network, complete restoration
small to stabilize the road networks in their present within five years, and new construction amounting
condition, let alone rehabilitate them. to at least 20 percent of the total.
Moderate to marginal financing capacity. Current
Country Differences funding would have to be increased by up to 50
percent and new construction held to 20 percent of
Countries differ widely in their financial require- the new total to ensure adequate maintenance and
ments for road maintenance and restoration, their complete the restoration in ten years. (Although
ability to marshal and allocate the needed resources, there is a wide definitional gap between this and
and their capacity to use additional resources effec- the preceding category, none of the countries fell
tively. within that gap; thus there is a clear distinction
between the two groups.)
Financial Constraints Insufficient financing capacity. Even if current
funding were increased by 50 percent and new
The financial severity of a country's road restoration construction held to 20 percent of the new total,
requirements can be gauged in two ways. One is to funds would be insufficient to complete the resto-
estimate how much the current funding for mainte- ration in ten years.
nance could be increased by reallocating funds within These criteria were applied to the sixty-one coun-
the overall budget for the sector. The other is to look tries for which expenditure data are available (see
at the rate of growth of real GNP per capita during table 4-4). More than half the countries with insuf-
the period 1975-85. The first criterion shows what ficient budgets to finance their road requirements
could be done with current road funds. The second (eleven of sixteen) had negative rates of real GNP
gives a crude indication of the potential for increas- growth. Of these, nine are in Sub-Saharan Africa,
ing those funds. Although any indicator of the poten- where extensive road building and paving in recent
tial for increasing funds is open to question, the decades set the stage for today's rapidly worsening
growth of real GNP per capita is believed to be a valid maintenance situation. These countries have the
indicator for most countries. (There are bound to be greatest need to increase their road budgets and the
exceptions: the oil-producing countries and others poorest prospects of doing so. They represent the
linked to them economically had unusually high hard core of the road maintenance problem in the
growth rates during 1975-85 but are now facing developing world.
adjustment problems because of the recent drop in A somewhat different problem affects China and
oil prices.) several countries in South Asia (Bangladesh, Burma,
Current road budgets (based on funding from India, and Pakistan). In general, many of their roads
domestic sources and, in some cases, external assis- were built long aogo to standards of strength and
tance) were placed in three broad categories accord- geometry that are totally inappropriate for today's

25
Table 4-4. Road Restoration aindMaintenance: Financing Capacity in Sixty-one Developing Countries

Annual growth
of real gross
national
product Capacity to finance restoration and maintenancea
per capita,
1975-85 Sufficient Marginal Insufficient

*Rep. of Korea {Botswana "People's Dem. Rep. of

Over 2.5 *Yemen Arab Rep. Arab Rep. of Egypt Yemen
percent Cameroon *Paraguay Burma
*Lesotho *Thailand Pakistan
*Indonesia Tunisia Sri Lanka
Algeria
Yugoslavia

Hungary India Bangladesh

0-2.5 *Colombia Morocco
percent Turkey Brazil
{'Rwanda Chile
*Burundi Portugal
'Oman "Burkina Faso
*Nepal "Benin
Honduras "Kenya
Ecuador Mali
*Malawi Guinea
Philippines
Dominican Rep.

*Papua New Guinea "Swaziland Sierra Leone

*Niger *Costa Rica "'Mauritania
Negative *C6te d'Ivoire Uruguay Zambia
growth *Nigeria Zimbabwe Tanzania
*Central African Rep. :Senegal
*Togo Madagascar
Argentina 'Gambia
"Liberia
Zaire
Jamaica
Bolivia

Notes: Countriesare listed in descendingorder of annualrate of growth of real GNP per capita for 1975-85.Countrieswith recentlyexpanded
or improvednetworksare markedwith an asterisk.
a. The capacityto financeis based on the nationalhighwaybudgets(1981-85),whichinclude,in certain cases,externalborrowingsor grants.
The assessmentof financialcapacityis somewhat subjectivebecausehighwaybudgetsfluctuatesignificantlyover time in many countries.This
classificationrhereforemay not accuratelyrepresentdevelopmentsafter 1985.

traffic. Although these roads, too, have deteriorated Reallocation of Funds

extensively, many of them should not be rehabilitat-
ed. Instead, they should be replaced by roads of a Aside from the countries with obsolete networks that
capacity adequate for present and expected require- need to be replaced, almost all the countries listed in
ments. table 4-4 will have to reallocate funds from new

26
construction to rehabilitation and maintenance if Above all, the managers and staff of the road agency
they cannot increase their road budgets significantly. will have to learn how to plan, manage, and super-
For many, the reallocations will have to be drastic. vise maintenance work efficientlyon a large scale.
Even countries with "sufficient" financing capacity Expansion may be especially troublesome for
will have to shift funds from one category to another, countries with recently expanded networks of paved
and this is difficult to do in many countries because roads. Many of these countries have had little expe-
of the budgeting process. Some countries have sepa- rience with maintenance and are ill prepared to carry
rate national budgets for capital and current expen- out the timely maintenance activities that will pre-
ditures, with construction in one and rehabilitation vent the need for far more expensive rehabilitation.
and maintenance in the other. In many countries road Slightlymore than half of the thirty countries in table
construction is part of the general development plan 4-4 with new paved networks are judged to have
and is controlled by a special office or ministry problems with their absorptive capacity. Although
separate from the ministry or bureau responsible for such judgments, which are based on Bank experi-
maintenance. In such a system it is almost impossible ence, may be subjective, the overall balance is realis-
to reallocate road funds from construction to reha- tic. Some Sub-Saharan countries with relatively new
bilitation and maintenance once the budget has been networks are affflictedby the worst combination
drawn up. (A change in allocations would require a of problems: insufficient financial capacity, nega-
top-level decision before budgets are submitted, and tive growth rates, and problematic institutional ca-
such a decision would probably encounter strong pacity.
resistance from the entity whose construction funds
were to be reduced.) Marshaling Domestic Resources
Reclassifying the budget could ease the conse-
quences of separating construction from rehabilita- Funds reallocated from new construction and gains
tion and maintenance. For example, resurfacing and in efficiencywill seldom be sufficient to finance the
rehabilitation of worn-out roads could reasonably be swelling costs of restoration and maintenance. A
classed as capital rather than recurrent expenditures. common method for increasing revenues to finance
Capital project planners might then pay more atten- road expenditures is to raise road user taxes and
tion to the need to maintain older roads and to assess other charges. Taxes on vehicles and supplies are
the future maintenance requirements of new con- often a convenient source of general revenue in
struction projects. The budgetary constraints on re- countries where the capacity to tax final products or
habilitation would also be less rigid. Such a reclas- incomes is limited. But high road user taxes may
sification should also be useful in reorienting distort transport costs and lead to economically inef-
the programs of external aid agencies. Another ficient business decisionsabout location, production,
possibility would be to integrate capital and current and investment. When distortions exist, taxes should
budgets and to select maintenance, rehabilitation, be raised only after their structure is adjusted to
and new construction projects on the basis of their reduce distortions. For example, where heavy vehi-
economic worth-taking into account future streams cles pay less than the cost of the damage they inflict
of maintenance requirements and road user savings on roads, a tax increase should be structured not only
in each case. to raise more revenue but also to encourage the use
of multiple axles to reduce road damage. To the
Absorptive Capacity extent that additional levies on road users are trans-
lated into better roads, they will reduce (not increase)
Money is not enough, because expanding the capac- road user costs.
ity to use funds effectivelycan be very difficult.Most Practically all the arguments of fiscaleconomicsgo
of the countries considered will have to expand their against earmarking taxes for specific purposes such
rehabilitation and maintenance operations, whether as road maintenance. Nevertheless, earmarking is
by shifting resources from construction or by other used in a variety of developing and developed coun-
means. Road maintenance agencieswill have to han- tries and should not be ruled out from consideration
dle maintenance efficiently and on a much larger in two cases. Where fiscal control is weak and reve-
scale than before. All this is likely to require a nue allocation and disbursement are subject to seep-
tremendous expansion of the maintenance capabili- age or long delay, earmarking can ensure that the
ties of the government or private contracting firms, government's decision to maintain and rehabilitate
or both. Machinery will have to be acquired, and roads is translated into practice. Short-circuiting the
operators and mechanics will have to be trained. budget may then make for a more rational use of

27
resources. In the second case-where a special tax is sively sharpened the focus of its lending for high-
to be levied (or a rate raised) to finance road reha- ways. From 1975 to 1985, 53 percent of the Bank's
bilitation-the principles of benefit taxation apply. lending for highways ($5.0 billion of $9.4 billion) was
Here the taxpayers know what they should expect to for rehabilitation, maintenance, and technical assis-
get in return. Such a targeted tax should have a tance largely related to maintenance. Some multila-
limited duration. teral and bilateral agencies, particularly those with a
In either case, earmarking can be recommended longer history of providing assistance for highway
only if the fiscal integrity of the receiving authority development, also raised the share of maintenance
(say, a road fund) and its determination to apply the and rehabilitation in their financing of roads. But
funds according to economic priorities are ensured. A from 1981 to 1984 most of them continued to provide
frequent problem is that money earmarked for the assistance for new construction, even to countries
road fund does not find its way there or is used with mounting maintenance backlogs. Bilateral aid
inefficiently or for purposes other than maintenance. tended to have a higher proportion of new construc-
One of the few reasonably successful schemes in tion and improvement works than did multilateral
developing countries emerged only after much trial assistance (table 4-5).
and error in the Central African Republic (see box In general, it makes sense to use external finance
4-1). primarily to increase a country's capital assets and
domestic finance to cover current costs, such as those
External Financing and Assistance for maintenance. (This is the tendency in most coun-
tries in East Asia and in Europe, the Middle East, and
The sheer size of the task now facing many develop- North Africa.) It also makes sense to regard the
ing countries implies massive demands for external willingness and ability of a country to pay for
financing. The foreign exchange components of the maintenance-an ongoing cost of road use-as a test
financing requirements range from 30 percent for of good internal management. There naturally are
routine maintenance in middle-income countries to exceptions, most obviously when countries have suf-
70 percent for restoration of paved roads in low- fered major calamities, when the structure of demand
income countries. In many cases external financing has changed considerably, or when the policies of
will have to cover more than the foreign exchange lenders are distorting factors. None of these excep-
component to preserve the infrastructure. In the tions will, however, justify lending for new construc-
group of countries at the hard core of the problem, tion if the existing roads are in serious disrepair. If
no solution is conceivable without external financing. donors and lenders restrict financing to the foreign
External financing will have to be concentrated on exchange cost of projects, and especially if they show
restoration and maintenance if the countries' efforts a strong preference for new construction, they tempt
are not to be diverted to lower priority investments recipient countries to divert their limited funds to
(Faiz and others 1987). The World Bank has progres- such projects as leverage to gain more foreign aid.

Box 4-1. Earmarked Road Funds in the Central African Republic

In the 1960s and 1970s the Central African Republic has the sole function of financing road maintenance.
made two unsuccessfulattempts to earmark funds for Revenuescome from a fuel tax, which the government
roads. The first road fund was subject to so much adjusts yearly to ensure the fund's capacity to do its
discretion that the earmarking had no practical effect, work. The fund is under the control of a ministerial
and the fund was abandoned after three years. The managementcommittee,on which road users are repre-
secondfund, set up in 1970in connectionwith the World sented through the Chamber of Commerce.
Bank's first road project, was designed to avoid the This third road fund has also had problems, but none
earlier shortcomings. But the amounts earmarked were fatal. At one time it had difficultycollectingthe revenues.
inadequate,payments were erratic, and more money had At another the government forced the fund to pay for a
to be sought through budgetary appropriations. project outside its mandate. Despite these and other
A third road fund, instituted in 1981 in connection problems the fund has improved road maintenance by
with the fourth highway project, is still functioning. As raising the level, and increasingthe regularity, of fund-
a public establishmentwith financialautonomy, the fund ing.

28
Table 4-5. External Assistance for Highways, 1981-84
Percentage distribution by major components

Millions New construction, Rehabilitation, Technical

Source of dollars improvements reconstruction Maintenance assistance,

World Bank 4,344 43 34 13 10

Other multilateral 2,889 n.a. n.a. n.a. n.a.
AfDB, ABEDA, AsDB, IDB,
IsDB, OPEC Fundb (1,892) 59 30 10 1i
Bilateral (DAC)d 1,529 n.a. n.a. n.a. n.a.
France, Federal Rep. of
Germany, and Japan (864) 62 25 8 5
Bilateral (other) 450 94 5 1

n.a. Not available.

.. Negligible.
a. Mostly for maintenance.
b. African Development Bank, Arab Bank for Economic Development in Africa, Asian Development Bank, Inter-American Development Bank,
Islamic Development Bank, Organization of the Petroleum Exporting Countries' Fund.
c. Self-standing technical assistance; project-related technical assistance is not included.
d. Development Assistance Committee.

Box 4-2. World Bank Lending in Chile

Until 1985 World Bank lending for roads in Chile was sive sector planning. The central road maintenance
mainly for individual construction projects. By 1985, budget was to be increasedfrom $78 million for 1986to
however, the rapidly increasing need for road mainte- $123million for 1988,with two-thirds of the investment
nance was already more than Chile could handle with its budget reallocated to rehabilitation and reconstruction
central funding and its decentralized institutional ar- ($44 million a year).
rangements. The municipal governments that were re- To support these moves and the 1986-88road invest-
sponsible for 70 percent (56,000kilometers) of the net- ment and maintenance program, the Bank provided a
work lacked the technical, managerial, and financial sector loan of $140 million (21 percent of the program).
resourcesneeded for the task. Disbursementsfrom the loan are tied to work on mainte-
On advice from the Bank the central road agency nance to ensure timely execution of this high-priority
(Vialidad)decided to resumegradually the responsibility program. The Bank also participated in a parallel $400
for local roads and to use private contractors wherever millionloan and worked withother lendersand suppliers
they were competitive.The Bank's Highway Design and to financethe rest of the program.
Maintenance Standards Model was used for comprehen-

Under these conditions it is not surprising that sometimes conflicting goals. As a first step toward
maintenance is neglected, especially in the poorest coordination the various lenders need to agree on
countries, where foreign funding plays a big part in a comprehensive road program with the recipient
determining the allocation of resources. Maintenance country. To ensure fiscal discipline in the long run,
work is then starved of the imported fuel, bitumen, all parties should scrutinize the implications of such
and spare parts needed for the efficient use of equip- a program: the capital and current budgets, the
ment and labor. In short, the incentives are biased relative priorities of different parts of the program,
against preventive maintenance and thereby contri- and the size of the program in relation to the re-
bute to the premature deterioration of assets. sources available from domestic and foreign sources.
To halt the deterioration of roads in developing This should make it easier to coordinate assistance
countries, all major lending agencies will have to to make the most effective use of resources.
liberalize their policies on lending for maintenance. Chile shows what such a coordinated program of
They will also have to coordinate their policies in highway sector lending can achieve (see box 4-2).
each country, instead of pursuing disparate and The requirements will differ in each country, so the

29
content and relative size of internal and external skills, and administrative capability) are greater than
contributions will also differ. For countries at the can be met without assistance.For these countries the
hard core of the road problem, the requirements in need for concerted action with donors is particularly
each dimension of the problem (money, technical great, since there is so little margin for waste.

30
 Conclusions and Policy
Recommendations

The conclusions and recommendations brought to- costly to repair, and this leads to further neglect and
gether in this chapter are based on three considera- premature failure.
tions: the increasing rate of road deterioration, the Good choicesin road managementdepend on many
insulation of road authorities from the consequences factors: climate, input prices, expected traffic flows,
of poor maintenance, and the huge costs and finan- vehicletypes and axle loads, existing road character-
cial requirements involved. istics and conditions, efficiencyof maintenance work,
The dynamics of road deterioration have much to available resources and the opportunity cost of capi-
do with the road maintenance problem. Paved roads tal, and attitudes toward differentrisks. Becausethese
do not deteriorate at a uniform rate. During an initial factors vary from place to place, universal prescrip-
phase of several years deterioration is minor, but it tions could lead to many more mistakes than success-
accelerates rapidly thereafter. Casual observation of es, and thus specific solutions are needed for each
road surfaces gives little warning of the imminence country. Nevertheless, some general conclusions
of the critical phase during which road conditions emerge from the empirical research of the past decade.
begin to deteriorate rapidly. If maintenance is ne-
glected, road users bear the brunt of the increase in Road planning and maintenance
total transport costs (the sum of infrastructure and * Gravel roads are more economical than paved
vehicle operating costs), since the share of agency roads where trafficvolumes are low, the climate is
costs in the total cost of transport is small. Al- not extreme, construction materials are not espe-
though vehicle operating costs constitute the domi- cially scarce, and adequate maintenance can be
nant share of the total costs of road transport, expected. This is true even for traffic volumes
improvements in the condition of uncongested substantially higher than the previously assumed
roads save less in vehicle operating costs than pre- limits for these roads. The uncertainty surround-
viously estimated. Project analysts therefore have ing some of the assumptions on which a decision
sometimes erred by crediting road investments with to pave has to be based allows for a broad range
greater benefits than are justified. Meanwhile, road of cost tradeoffs in which the break-even traffic
authorities have often erred by ignoring the effect of volume for paving varies from less than 100 vehi-
neglected maintenance on user costs. For road pav- cles a day to more than 400, depending on the
ing, moreover, the two errors do not cancel each tradeoff. Decisions should therefore be guided by
other out; they reinforce one another. Paving is done prudence and be based on a careful assessment of
before it is warranted, and then the pavements are local conditions and of the likelihood that the road
neglected. The pavements become more and more will indeed be adequately maintained.

31
* According to present knowledge, some paved Institutional development
roads should have been left unpaved because they Inadequate maintenance in developing countries
carry so little traffic. Rather than maintain such has various causes, but only institutional failure
roads at normal standards, it may be economical, can explain the extent of the inadequacy. At the
when funds are limited, to let them deteriorate and heart of this failure is the absence of public ac-
simply use low-cost patching to keep them usable. countability. All activities to strengthen institu-
* If a road carrying more than about 500 vehicles a tions, enhance incentives, and improve the internal
day is to be paved or strengthened, and if axle workings of road agencies should be judged by
loads are hard to control, the economic savings their ability to increase accountability. The road
from staged construction of the pavement are agency itself should be subject to an independent
likely to be less than the cost of premature pave- system of auditing and inspection.
ment failure. Thus countries that have difficulty The nature of road deterioration and the separa-
enforcing load limits should build roads to higher tion of road management from road use have
initial standards, even though this normally means sheltered road authorities from public pressures
building fewer roads. and market signals. Both user groups and highway
* Because of the nonlinear pattern of deterioration administrations should stimulate public awareness
of paved roads, the resurfacing of newer roads can and communicate their needs and problems to
be deferred with only a small penalty-but only if policymakers.
those roads are not too close to their critical age and * To limit the potential for conflict between the
not heavily traveled. If the roads are somewhat planning and control function of the road author-
older or traffic is heavier, the deferral of resurfacing ity and the work execution function, the two
could cause the pavements to break apart and make should be separated. The government should min-
costly reconstruction necessarylater on. imize its direct role in work execution, even for
* In ranking the factors affecting maintenance routine maintenance, and transfer that responsibil-
choices,traffic volume is generally more important ity to independent entities operating in a compet-
than road condition. So when budgets are severely itive environment that encourages managerial flex-
constrained, across-the-board cutbacks may not be ibility and efficiency.
the answer. It may be better to maintain high-
volume roads in fair or good condition and to Financing
reduce substantially the maintenance of some low- * In countries with a backlog of roads needing
volume roads. rehabilitation and in those with young roads ap-
* Because the deterioration of paved roads is insidi- proaching the age when maintenance requirements
ous, a road maintenance agency must regularly multiply, adequate resources should be brought to
monitor the condition of roads so as not to delay bear on the maintenance problem before the roads
maintenance beyond the point at which costs rise get worse. This focus often requires reallocating
steeply. The agency must also monitor the volume funds from new construction to rehabilitation and
and mix of traffic and axle loads to determine maintenance. Ways of increasing the total road
priorities for investment and maintenance. An ef- budget should be sought; increased user fees and
fective management information system that cov- taxes, possibly earmarked for rehabilitation and
ers these conditions as well as the agency's maintenance, should be considered.For many coun-
equipment, supplies, and personnel is a basic re- tries, external financing will also be needed-with-
quirement for adequate planning and deployment out it the serviceablenetwork will have to contract.
of resources. * In many countries road allocations will have to be
* For maximum economy, road maintenance policy transferred from new construction to maintenance
should be coordinated with road design and con- unless the total budget can be increased. Where
struction planning, and the life-cyclecosts of the budgets for the two purposes are separate or where
roads should be balanced against the operating a single authority cannot reallocate funds with-
costs of the vehiclesusing them. This coordination in the total, road resurfacing and rehabilitation
requires a good data base, capable staff, and sound should be classed as a capital expenditure.
analytical techniques. Appropriate models and in-
formation managementsystemsexist for such anal- WorldBank Policy
yses. Their use will not prevent errors that result
from bad data or incorrect traffic forecastsbut may Recommendations for World Bank policy must dis-
help to avoid the errors of oversimplification. tinguish between principles that apply to all highway

32
 Box 5-1. A Diagnostic Framework for Determining External Assistance Policy in the Road Sector

The country examplesused in this diagnosticframework mobilize substantial external financial and institutional
represent conditions in 1984. assistance.They must give high priority to the coordina-
tion of aid programs and emphasizeinstitutional reform.
Category 1. In these countries the institutional capacity A. This subcategory of countries with relativelynew
and past maintenance effort have generally been ade- networks must give priority to resurfacingand strength-
quate. These countries now need to adjust policies and ening paved and unpaved roads and to building institu-
expenditures to meet emerging maintenance require- tional capacity.Examples: Benin,Burundi, Liberia,Mali,
ments, particularly for new networks. They also need to Nepal, Senegal.
continue to make institutional and technological im- B. This subcategory of countries must give priority to
provements and find ways to increase operational ef- restoring their aging road infrastructure and to building
ficiency.Examples: Chile, Republic of Korea, Malawi, the institutional capacity in the road sector to aid eco-
Niger, YemenArab Republic. nomic recovery.Examples: Bolivia,Ghana, Laos, Mad-
agascar, Sierra Leone, Tanzania.
Category II. In these countries the funding for mainte-
nance has been insufficient.They urgently need to real- Category IV. In these countries, the networks and
locate domesticand external resourceswithin and to the maintenance technology are obsolete. They need to
road sector. They also need to devote greater emphasis mobilizedomestic and external resourcesfor moderniza-
to policy reform to expand their institutional capacity. tion. They also need to place emphasis on technology
Examples:Brazil,Indonesia, Kenya,Nigeria, Yugoslavia. transfer, institutional improvements, and the develop-
ment of skills. Examples: Bangladesh, Burma, China,
Category 111.In these countries the institutional capacity India, Pakistan.
has been grossly inadequate. They urgently need to

lending and principles that apply to different coun- serious road transport constraints over the next
tries according to rheir capacity for financing estimat- five to ten years. The main emphasis of external
ed maintenance needs and the capacity of their insti- assistance should be on institutional improve-
tutions. A review of information, backed by the ments, technology transfer, and operational ef-
experience of Bank staff, suggests that countries can ficiency.
be placed in four main classes according to: Category II countries have underfunded road
* The state of the road networks maintenance, with the result that it has been inad-
* The financial requirements for maintenance and equate in quality or quantity, or both. They could,
rehabilitation and the country's ability to meet however, rectify past inadequacies over five to ten
those requirements years by reallocating resources to maintenance,
* The possibility of reallocating recurrent funds for improving the efficiency of their operations and
roads without causing severe disruptions management, securing more external assistance,
* The institutional capacity to absorb more funding and, in some cases, increasing the road budget.
for maintenance and rehabilitation. Category III countries have backlogs of rehabilita-
The four categories are reasonably exhaustive and tion and maintenance so large relative to domestic
cover cases that the Bank is likely to encounter in the funds, personnel, and technology that corrective
next few years. Each category can be exemplified by action will take substantial time, effort, and exter-
countries for which adequate information is available nal assistance. This category is divided into coun-
(see box 5-1). tries in which preserving a young network is the
* Category I countries have, for the most part, used primary need (IIIA) and countries with older net-
resources for road investments and maintenance in works needing massive rehabilitation (IIIB).
a cost-effective manner and now have a serviceable Category IV includes China and most of the Indian
road network. Sustaining past trends will depend subcontinent. The common characteristic is an
on the ability of these countries to adjust their road extensive and obsolete road network in need of
priorities to changing circumstances and to main- modernization to meet the burgeoning demand for
tain their existing institutional capacity. Provided road transport. With few exceptions, these coun-
these conditions are met and barring unforeseen tries have not neglected maintenance. But their
events, countries in this category should suffer no aged road networks and outmoded maintenance

33
 technology seriously constrain good maintenance tranches and made conditional on reaching the
practice. These countries need to mobilize substan- interim objectives of agreed-on programs.
tial domestic and external financial resources to * In Bank appraisals of road construction and bet-
meet the growing needs of their road networks. terment, the probability of a road's being main-
A common requirement for all four categories, and tained to acceptable standards will enter explicitly
one of great importance to Bank policy, is the real- into the calculation of a project's expected net
location of domestic and external resources to bal- benefitsand will be determined in large part by the
ance new construction with maintenance, reha- country's past record of road maintenance.
bilitation, and betterment. Reallocation will vary Because so much capital is at risk and the eco-
according to each country's circumstanceswithin the nomic outcomes are so sensitive to country differ-
broad limits for each category. In most countries, ences in networks, traffic, and other circumstances,
reallocating the road budget will leave fewer re- the Bank will require that road agencies have an
sources for new construction, so that new investment adequate road management system-or a phased,
priorities will have to be established. In some coun- monitorable program for establishing one. To im-
tries, moreover, no amount of reallocation will suf- prove efficiencyand reduce the potential for conflict
fice; nor will the economic situation permit additions between the planning and execution functions of the
to the road budget. In such circumstances the net- road authority, the Bank will encourage maintenance
work's quality and extent will have to be reduced. If work by (private or public) entities outside the road
such a contraction appears inescapable,the objective authority that operate according to commercial prin-
of the Bank's analytic work will be to establish ciples and, preferably, in a competitive environment.
priorities for essential maintenance and restoration. In addition, the Bank will support schemes to in-
The Bank's sector and economic work in transport crease awareness of the need for management sys-
will refine the assignment of countries to the four tems and for adequate resources for rehabilitation
categories. Moreover, the analysis of a country's road and maintenance. These schemeswill focus on efforts
sector will be essential to formulating the Bank's to educate government officials and to make the
recommendations, particularly for countries having public aware of the consequences of neglecting
the most critical problems. Because restoration and maintenance. Such efforts will include regular publi-
maintenance needs vastly exceed the resources avail- cation of the data obtained from monitoring road
able for this purpose in most countries, public invest- conditions.
ment and expenditure reviews will recommend the The Bank will also assessprogress in the education
division of road expenditures among new construc- and public awareness components of Bank projects
tion, restoration, and maintenance. These reviews and in the transfer of maintenance work to entities
will also evaluate the financial and institutional ca- outside the road agency. It shall periodically re-
pacity for coping with the future maintenance needs view the four-category classification of countries
of new road investments and propose cost-effective and make appropriate adjustments on the basis of
investment and maintenance strategies. new evidence or the findings of its operational staff.
Bank lending for roads will be conditional on each
country's achievement of an acceptable distribution Action by the International Community
of expenditures among new construction, restora-
tion, and maintenance. In financing highways, external agencies should em-
* For countries in category III, lending will normally phasize rehabilitation and maintenance in countries
be confined to road maintenance and restoration in categories II and III and modernization in coun-
and be made contingent upon the country's apply- tries in category IV. They should limit their financing
ing its entire road budget, including external assis- of new construction largely to countries in categories
tance, to maintenance and restoration. Exceptions I and IV-and even then provide funds only through
to this condition will be made where roads are integrated programs for construction, maintenance,
needed to exploit agricultural and mineral re- rehabilitation, and modernization. In addition, all
sources in the context of identifiable investment external agencies involved in road sector activitiesin
projects. any country in category II or III should require the
* For countries in category II, lending will be condi- recipient government to establish internal adminis-
tional on periodic agreement between the govern- trative mechanisms to coordinate and monitor ex-
ment and the Bank about the division of funds ternal assistance programs. Consultative group or
among new construction, restoration, and mainte- round table meetings can effectively help govern-
nance. Funds will be committed or released in ments to achieve this objective.

34
 The data base needed for identifying the factors international statistics on road conditions. Such a
that influence road deterioration and the effects of compilation is essential if worldwide trends are to be
different maintenance processes should be substan- detected and reliable judgments made about the
tially strengthened and refined. Road research insti- relative performance of countries. The United Na-
tutions and highway authorities in developed and tions Statistical Office-or an organization such as
developingcountries should continue research on the the Permanent International Association of Road
multidimensional issues of road deterioration-tech- Congresses or the International Road Federation-
nical, institutional and financial-and should sup- could assume responsibility for compiling these sta-
port international exchanges of data and technology. tistics. Also recommended is the annual publication
In addition, external agenciesshould provide techni- of statistics on the sources and uses of external aid
cal and financial support for the compilation of for road development.

35
 APPENDIX

StatisticalTables

A-1 Basic Characteristics of Road Networks by Region 39

A-2 Road Networks by Country and Region, 1984 40

37
 Table A-1. Basic Characteristics of Road Networks by Region

East Asia Europe, the Latin America

Eastern and WesternAfrica and the Pacific South Asia Middle East, and the Caribbean
Southern Excluding Excluding Excluding and Excluding Brazil
item Africa Ali Nigeria All China Ali India North Africa All and Mexico Total
Numberofcountries 19 20 19 7 6 6 5 14 19 17 85
Area(thousandsofsquarekilometers) 11,051 8,891 7,967 13,184 3,623 5,120 1,922 6,476 18,859 8,374 63,581
GNP,1984 (billions of dollars) 51.2 104.7 34.3 600.4 281.4 257.5 62.7 338.1 583.9 199.1 1,935.9
Population, 1984 (millions) 201 186 90 1,350 221 1,008 259 232 350 141 3,327
GNP per capita
Median, 1984(dollars) 310 290 270 710 785 220 180 1,795 1,190 1,160 -
Average annual growth rate,
1975-85 (percent) -0.1 -1.0 -0.9 3.4 2.5 2.4 1.8 3.1 0.0 -0.1 -
Total road network
Length(thousandsofkilometers) 588.3 430.9 322.9 1,539.6 571.2 1,680.7 180.7 1,051.0 2,212.1 874.1 7,502.6
Percentage paved 7.7 12.6 10.3 18.1 16.9 31.1 26.7 29.3 11.0 11.8 19.4
Density in kilometers per:
lOOsquarekilometers 5.3 4.8 4.1 11.7 15.8 32.8 9.4 16.2 11.7 10.4 11.8
1,00Opopulation 2.9 2.3 3.6 1.1 2.6 1.7 0.7 4.5 6.3 6.2 2.3
$1 million of GNP 11.5 4.1 9.4 2.6 2.0 6.5 2.9 3.1 3.8 4.4 3.9
Main roads
Length(thousandsofkilometers) 190.5 144.5 115.4 450.9 196.6 217.1 89.6 303.7 511.7 193.7 1,818.4
Percentage paved 24.6 37.3 27.8 61.4 52.8 80.3 56.0 74.5 49.2 44.3 56.8
Density in kilometers per:
100 square kilometers 1.7 1.6 1.4 3.4 5.4 4.2 4.7 4.7 2.7 2.3 2.9
1,000 population 0.9 0.8 1.3 0.3 0.9 0.2 0.3 1.3 1.5 1.4 0.5
$1 million of GNP 3.7 1.4 3.4 0.8 0.7 0.8 1.4 0.9 0.9 1.0 0.9
Main paved roads
Length(thousandsofkilometers) 46.9 53.9 32.1 276.9 103.8 174.3 50.2 226.3 251.8 85.8 1,030.0
Density (kilometers per 100 square
kilometers) 0.8 0.9 0.5 9.5 5.6 7.4 2.3 10.6 6.8 2.0 29.5
Replacement value of main network
Average cost per kilometer (dollars)
Paved 255,000 245,000 245,000 300,000 300,000 180,000 180,000 280,000 260,000 260,000 -
Unpaved 40,000 40,000 40,000 40,000 40,000 40,000 40,000 40,000 40,000 40,000 -
Value (billions of dollars)
Paved 12.0 13.2 7.9 83.1 31.1 31.4 9.0 63.4 65.5 22.3 268.4
Unpaved 5.7 3.6 3.3 7.0 3.7 1.7 1.6 3.1 10.4 4.3 31.5
Total 17.7 16.8 11.2 90.0 34.9 33.1 10.6 66.4 75.9 26.6 299.9
As a percentage ofGNP 34.6 16.1 32.6 15.0 12.4 12.9 16.9 19.7 13.0 13.4 15.5
- Not applicable.
Source: World Bank surveys and reports.

CA)
CD
Table A-2. Road Networks by Country and Region, 1984
GNPper Capita Total network Ma
A rea Average Density in kilometers per:
(thousands annualgrowth Total GNP 100 $1
Country Population of square 1984 rate, 1975-85 (millions Length square 1,000 million Length
and region (millions) kilometers) (dollars) (percent) of dollars) (kilometers) kilometers populatiot Of GNP (kilometer
Eastern and
Southem Africa
Botswana 1.0 600 960 7.9 960 13,00(0 2.2 13.0 13.5 8,026
Burundi 4.6 28 220 0.9 1,012 5,400 19.3 1.2 5.3 2,900
Comoros 0.4 2 340 n.a. 129 85() 42.5 2.2 6.6 762
Djibouti 0.4 22 358 n.a. 129 2,800 12.7 7.8 21.7 1,100
Ethiopia 42.2 1,222 110 -0.9 4,642 43,200 3.5 1.0 9.3 13,600
Kenya 19.6 583 310 0.1 6,076 55,000 9.4 2.8 9.1 25,300
Lesotho 1.5 30 530 3.7 795 4,00(0 13.3 2.7 5.0 2,010
Madagascar 9.9 587 260 -3.4 2,574 50,000 8.5 5.1 19.4 10,15()
Malawi 6.8 118 180 0.2 1,224 13,280 11.3 2.0 1().8 5,571
Mauritius 1.1 2 1,090 2.6 1,199 2,795 139.8 2.5 2.3 1,787
Rwanda 5.8 26 280 1.( 1,624 10,000 38.5 1.7 6.2 4,000
Somalia 5.2 638 260 -0.3 1,352 21,600 3.4 4.2 16.0 9,644
Sudan 21.3 2,506 360 -0.4 7,668 19,110 0.8 0.9 2.5 7,000
Swaziland 0.7 17 80( -0.20 568 2,821 16.6 4.0 5.0 2,724
Tanzania 21.5 945 210 -1.7 4,515 53,600 5.7 2.5 11.9 17,70(0
Uganda 15.0 236 230 -3.3 3,450 27,037 11.5 1.8 7.8 5,691
Zaire 29.7 2,345 140 -3.9 4,158 145,000 6.2 4.9 34.9 4(0,900
Zambia 6.4 753 470 -3.7 3,008 35,00(0 4.6 5.5 11.6 20,653
Zimbabwe 8.1 391 760 -1.3 6,156 85,000 21.7 10.5 13.8 11,003
Total 201.2 11,051 - -- 51,239 589,493 5.3 2.9 11.5 190,521
Average - - 255 -0.1 - - - - - -
WesternAfrica
Benin 3.9 113 270 0.6 1,053 7,500 6.6 1.9 7.1 4,926
BurkinaFaso 6.6 274 160 0.9 1,056 11,200 4.1 1.7 10.6 6,3(0(0
Cameroon 9.9 475 760 3.9 7,524 64,905 13.7 6.6 8.6 13,50(0
Central African Rep. 2.5 623 260 -1.4 650 22,600 3.6 9.( 34.8 5,25(0
Chad 4.9 1,284 80 _7.7b 392 31,300 2.4 6.4 79.8 3,8()0
Congo 1.8 342 1,140 2.2 2,052 11,000 3.2 6.1 5.4 7,00(0
C6ted'lvoire 9.9 322 610 -2.2 6,039 47,880 14.9 4.8 7.9 13,680
Equatorial Guinea 0.4 28 180 n.a. 67 1,540 5.5 4.2 23.1 1,(9(
Gambia,The 0.7 11 260 -1.7 185 2,388 21.7 3.4 12.9 1,209
Ghana 12.3 239 350 -1.9 4,305 28,400 11.9 2.3 6.6 14,130
Guinea 5.9 246 330 0.2h 1,947 14,00(1 5.7 2.4 7.2 4,550
Guinea-Bissau 0.9 36 180 -2.11 158 4,040 11.2 4.6 25.5 2,300
Liberia 2.1 111 470 -3.7 987 7,560 6.8 3.6 7.7 3,945
Mali 7.3 1,240 140 0.0 1,022 13,500 1.1 1.8 13.2 5,222
Mauritania 1.7 1,031 450 -0.4 765 7,500 0.7 4.4 9.8 2,48(0
Niger 6.2 1,267 190 -1.1 1,178 19,0()0 1.5 3.1 16.1 6,369
Nigeria 96.5 924 730 -3.3 70,445 108,000 11.7 1.1 1.5 29,1(0(0
Senegal 6.4 196 480 -1.3 3,072 11,700 6.0 1.8 3.8 9,7(0(0
Sierra Leone 3.7 72 310 -0.3 1,147 9,924 13.8 2.7 8.7 7,04(0
Togo 2.9 57 250 - 1.5 725 7,000 12.3 2.4 9.7 2,921
Total 186.5 8,891 - - 10)4,769 430,937 4.8 2.3 4.1 144,512
Average - - 562 - 1.0 - - - - - -
East Asia and
the Pacific
China 1,029.2 9,561 310 6.8 319,052 915,100 9.6 (1.9 2.9 254,30(0
Indonesia 158.9 1,919 540 3.6 85,806 201,30(0 10.5 1.3 2.3 45,80(0
Korea, Rep. of 40.1 98 2,110 6.4 84,611 54,00(1 55.1 1.3 (0.6 23,481
Malaysia 15.3 330 1,980 4.1 30,294 45,0(0(0 13.6 2.9 1.5 25,324
PapuaNewGuinea 3.4 462 710 -0.8 2,414 18,545 4.0 5.5 7.7 3,652
Philippines 53.4 300 660 0.1 35,244 155,669 51.9 2.9 4.4 53,864
Thailand 50.0 514 860 3.9 43,000 150,000 29.2 .3.( 3.5 44,534
Total 1,350.3 13,184 - - 600,421 1,539,614 11.7 1.1 2.6 4510,955
Average - - 445 3.4 - - - - - -
South Asia
Bangladesh 98.1 144 130 1.7 12,753 15,()0() 111.4 (1.2 1.2 111,9110
Burma 36.1 677 180 3.9 6,498 27,983 4.1 0.8 4.3 25,40((

40
network Main paved network Main unpaved network
Density in kilometers per: Portion constructed
100 $1 Condition Portionconstructed orimproved n: Condition or improved
square 1,000 million Length (percent)' 1980-84 1975-79 Length (percent)' in 1980-84
kilometers population ofGNP (kilometers) Good Fair Poor Kilomneters Percent Kilometers Percent (kilometers) Good Fair Poor Kilometers Percent

1.3 8.0 8.4 1,967 90 8 2 600 31 400 20 6,059 40 30 30 400 7

10.4 0.6 2.9 744 58 37 5 390 52 40 5 2,156 60 25 15 n.a. n.a.
38.1 2.0 5.9 456 60 20 20 50 11 21 5 306 20 20 60 n.a. n.a.
5.0 3.1 n.a. 300 30 60 10 24 8 n.a. n.a. 800 10 10( 80 na. n.a.
1.1 0.3 2.9 3,800 33 34 33 485 13 n.a. n.a. 9,800 33 34 33 n.a. n.a.
4.3 1.3 4.2 6,000 32 52 16 1,400 23 1,600 27 19,300 66 17 17 358 2
6.7 1.3 2.5 398 25 35 40 n.a. n.a. n.a. n.a. 1,612 20 50 30 n.a. n.a.
1.7 1.0 3.9 4,890 20 30 50 300 6 450 9 S,260 20 20 60 n.a. n.a.
4.7 0.8 4.6 2,004 50 35 15 270 13 913 46 3,567 30 35 335 400 11
89.4 1.6 I.S 1,610 n.a. n.a. n.a. n.a. n.a. n.a. n.a. 177 n.a. n.a. n.a. n.a. n.a.
15.4 0.7 2.5 800 80 20 0 300 38 60 8 3,200 20 80 0 3,00() 94
1.5 1.9 7.1 2,460 60 20 20 n.a. n.a. n.a. n.a. 7,184 30 30 40 n.a. n.a.
0.3 0.3 0.9 2,240 40 30 30 790 35 n.a. n.a. 4,760 20 20 60 n.a. n.a.
16.0 3.8 4.8 521 50 25 25 285 55 80 15 2,203 25 25 50 n.a. n.a.
1.9 0.8 3.9 3,240 38 34 28 581 18 n.a. n.a. 14,460 5 47 48 1,700 12
2.4 0.4 1.6 1,726 10 59 31 n.a. n.a. 110 6 3,965 52 26 22 n.a. n.a.
1.7 1.4 9.8 2,400 17 20 63 200 8 400 17 38,500 61 25 14 n.a. n.a.
2.7 3.2 6.9 5,502 40 30 30 150 3 1,200 22 15,151 30 35 35 25(0 2
2.8 1.4 1.8 5,595 70 25 5 n.a. n.a. n.a. n.a. 5,408 50 3(0 2(1 2,090 39
- - - 46,653 - - - - - - - 143,868 - - - - -
1.7 0.9 3.7 - 42 32 26 - - - - - 42 30 28 - -

4.4 1.3 4.7 986 58 36 6 230 23 100 10 3,940 15 55 30 500 13

2.3 1.0 6.0 1,400 40 40 20 400 29 281 20 4,900 1() 30 60 3 55 7
2.8 1.4 1.8 2,900 60 30 10 783 27 797 27 1(0,600 20 50 30 n.a. n.a.
0.8 2.1 8.1 442 36 35 29 269 61 65 15 4,808 68 16 16 955 20
0.3 0.8 9.7 163 4 0 96 7 4 n.a. n.a. 3,637 8 19 73 1,(00 27
2.0 3.9 3.4 1,100 69 15 16 620 56 480 44 5,900 30 2(1 50 150 3
4.2 1.4 2.3 3,620 78 15 7 955 26 1,330 37 10,060 3(0 6() 10 170 2
3.9 2.9 n.a. 485 26 0 74 n.a. n.a. 80 16 605 0 ( 100 n.a. n.a.
11.0 1.7 6.5 447 27 50 23 210 47 30 7 762 30 42 28 310 41
5.9 1.1 3.3 5,782 12 27 61 126 2 n.a. n.a. 8,348 17 48 35 n.a. n.a.
1.8 0.8 2.3 1,145 51 35 14 201 18 85 7 3,405 13 22 65 975 29
6.4 2.6 14.5 485 2 65 33 30 6 30 6 1,815 n.a. n.a. n.a. n.a. n.a.
3.6 1.9 4.0 557 85 13 2 110 20 250 45 3.388 15 75 10 176 5
0.4 0.7 5.1 1,890 43 33 24 540 29 378 20 3,332 11 4 85 1,340 4(1
0.2 1.5 3.2 1,640 30 30 40 1,200 73 200 12 840 0 5 95 n.a. n.a.
0.5 1.0 5.4 2,609 70 20 10 505 19 757 29 3,76(0 30 35 35 666 18
3.1 0.3 0.4 21,100 62 15 23 4,980 24 7,588 36 8,000 0 1(0 90 n.a. n.a.
4.9 1.5 3.2 3,762 51 36 13 470 12 740 20 5,938 4 26 7(0 400 7
9.8 1.9 6.1 1,280 20 45 35 240 19 n.a. n.a. 5,760 30 40 30 n.a. n.a.
5.1 1.0 4.0 1,712 40 24 36 427 25 235 14 1,209 30 45 25 86 7
- - - 53,505 - - - - - - - 91,0(07 - - - - -
1.6 0.8 1.4 - 52 23 25 - - - - - 20 36 44 - -

2.7 0.2 0.8 173,000 10 70 20 22,000 13 59,000 34 81,300 50 25 25 n.a. n.a.

2.4 0.3 0.5 27,550 30 30 40 5,530 20 5,000 18 18,250 22 47 31 3,5(00 19
24.0 0.6 0.3 9,928 70 25 5 3,860 39 1,926 19 13,553 60 20 20 n.a. n.a.
7.7 1.7 0.8 23,479 n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1,845 n.a. n.a. n.a. n.a. n.a.
0.8 1.1 1.5 726 34 45 21 110 15 109 15 2,926 23 71 6 n.a. n.a.
18.0 1.0 1.5 14,192 31 54 15 900 6 500 4 39,672 25 50 25 I,(55 3
8.7 0.9 1.0 28,000 50 30 20 8,6(00 31 5,242 19 16,534 45 30 25 n.a. n a.
- - - 276,875 - - - - - - - 174,080 - - - - -
3.4 0.3 0.8 - 20 59 21 - - - - - 41 34 25 - -

7.6 0.1 0.9 6,210 15 40 45 1,180 19 60 1 4,690 10 4(0 5( n.a. n.a.

3.8 0.7 3.9 8,400 0 5() 50 n.a. n.a. n.a. n.a. 17,0()0 0 50 50 n.a. n.a.
(Table continues on the following page.)

41
TableA-2. (continued)
GNP per capita -lbtal network Main
Area Average Density in kilometers per:
(thousands annualgrowth Total GNP 100 Sl
Country Population of square 1984 rate, 1975-85 (millions Length square 1,000 million Length
and region (millions) kilometers) (dollars) (percent) of dollars) (kilometers) kilometers population (of GNP (kilometers)
India (national) 749.2 3,288 260 1.9 194,792 1,500,000 45.6 2.0 7.7 32,000
India (states) - - - - - - - - - 95,500
Nepal 16.1 141 160 0.7 2,576 7,150 5.1 0.4 2.8 5,546
Pakistan 92.4 804 380 3.3 35,112 107,673 13.4 1.2 3.1 38,830
Sri Lanka 1S.9 66 360 3.2 5,724 25,500 38.6 1.6 4.5 8,900
Total 1,007.8 5,120 - - 257,455 1,683,306 32.9 1.7 6.5 217,076
Average - - 255 2.4 - - - - - _
Europe, the
Middle East,
and North Africa
Algeria 21.2 2,382 2,410 3.2 51,092 78,19(0 3.3 3.7 1.5 39,347
Cyprus 0.7 9 3,590 7.0 2,405 6,831 75.9 10.2 2.8 3,031
Egypt, Arab Rep. of 45.9 1,001 720 4.8 33,048 30,089 3.0 0.7 0.9 28,725
Hungary 10.7 93 2,100 5.lb 22,470 138,185 148.6 12.9 6.1 24,000
Morocco 21.4 447 670 1.4 14,338 57,692 12.9 2.7 4.0 19,080
Oman 1.1 300 6,490 4.4" 7,139 18,123 6.0 16.5 2.5 8,474
Portugal 10.2 92 1,970 1.6 20,094 52,031 56.6 5.1 2.6 19,031
Romania 22.7 238 2,290 n.a. 51,983 73,500 30.9 3.2 1.4 14,7()0
Syria 10.1 185 1,620 -0.2 16,362 26,200 14.2 2.6 1.6 15,700
Tunisia 7.0 164 1,270 2.6 8,890 78,190 47.7 11.2 8.8 16,900
Turkey 48.4 781 1,160 1.4 56,144 329,793 42.2 6.8 5.9 60,953
Yemen Arab Rep. 7.8 195 550 4.0 4,290 25,028 12.8 3.2 5.8 3,028
Yemen, People's
Dem. Rep. of 2.0 333 550 5.2 1,100 7,100 2.1 3.6 6.5 1,887
Yugoslavia 23.0 256 2,120 2.6 48,760 130,000 50.8 5.7 2.7 48,880
Total 232.2 6,476 - - 338,115 1,050,952 16.2 4.5 3.1 303,736
Average - - 1,456 3.1 - - - - -

Latin America and

the Caribbean
Argentina 30.1 2,767 2,230 -1.9 67,123 212,305 7.7 7.1 3.2 36,5(05
Barbados 0.3 0 4,340 1.6 1,128 1,670 388.4 6.4 1.5 1,367
Belize 0.2 23 1,150 3.6" 184 2,000 8.7 12.5 10.9 1,871
Bolivia 6.2 1,099 540 -4.0 3,348 39,824 3.6 6.4 11.9 9.382
Brazil (federal) 132.6 8,512 1,720 1.2 228,072 1,316,343 15.5 9.9 5.8 64,642
Brazil (states) - - - - - - - 162,596
Chile 11.8 757 1,700 0.9 20,060 79,000 10.4 6.7 3.9 22,831
Colombia 28.4 1,139 1,390 1.7 39,476 77,200 6.8 2.7 2.0 24,397
CostaRica 2.5 51 1,190 --0.3 2,975 28,500 55.9 11.4 9.6 4,789
Dominican Rep. 6.1 49 970 0.2 5,917 17,200 35.1 2.8 2.9 3,475
Ecuador 9.1 284 1,150 0.3 10,465 37,910 13.3 4.2 3.6 8,400
Guatemala 7.7 109 1,160 --.3.3 8,932 18,426 16.9 2.4 2.1 10,641
Haiti 5.4 28 320 0.7 1,728 4,000 14.3 0.7 2.3 2,473
Honduras 4.2 112 700 0.4 2,940 12,058 10.8 2.9 4.1 6,018
Jamaica 2.2 11 1,lS0 -4.0 2,530 17,700 160.9 8.( 7.0 4,750
Mexico (federal) 76.8 1,973 2,040 0.6 156,672 212,626 10.8 2.8 1.4 43,500
Mexico (states) - - - - - - - - - 47,300
Panama 2.1 77 1,980 1.3 4,158 8,665 11.3 4.1 2.1 8,4(0(0
Paraguay 3.3 407 1,240 3.7 4,092 24,329 6.0 7.4 5.9 11,241
Peru 18.2 1,285 1,000 -2.5 18,200 58,516 4.6 3.2 3.2 27,3001
Uruguay 3.0 176 1,980 -1.0 5,940 49,813 28.3 16.6 8.4 9,81.3
Total 350.2 18,859 - - 583,940 2,218,085 11.8 6.3 3.8 511,691
Average - - 1,668 -0.03 - - - - - -
Total for all six regions
(85 countries) 3,328.2 63,581 - - 1,935,939 7,512,387 - - - 1,818,491
Average - - 582 - - - - - ---

-Not applicable.
n.a. Not available.
a. Percentage of network in each category.
b. 1973-83.
Source: World Bank surveys and reports.

42
network Main paved network Main unpaved network
Density in kilometers per: Portion constructed
100 $1 Condition Portion constructed or improved in: Condition or improved
square 1,000 ml Length (percent)' 1980-84 1975-79 Length (percent) in 1980-84
kilometers population of GNP (kilometers) Good Fair Poor Kilometers Percent Kilometers Percent (kilometers) Good Fair Poor Kilometers Percent
1.0 0.0 0.2 31,600 20 45 35 2,500 8 3,000 9 400 20 40 40 n.a. n.a.
- - - 91,200 20 45 35 n.a. n.a. n.a. n.a. 4,300 20 40 40 n.a. n.a.
3.9 0.3 2.2 2,500 40 35 25 584 23 376 15 3,046 15 45 40 300 10
4.8 0.4 1.1 24,142 18 50 32 1,940 8 2,500 10 14,688 5 25 70 n.a. n.a.
13.5 0.6 1.6 8,900 10 40 50 50 1 100 1 0 0 0 0 n.a. n.a.
- - - 172,952 - - - - - - - 44,124 - - - - -
4.2 0.2 0.8 - 19 45 36 - - - - - 6 39 55 - -

1.7 1.9 0.8 32,480 40 32 28 n.a. n.a. 1,070 3 6,867 25 50 25 n.a. n.a.
33.7 4.5 1.3 2,938 38 38 24 253 9 400 14 93 0 50 50 n.a. n.a.
2.9 0.6 0.9 15,081 39 33 28 n.a. n.a. n.a. n.a. 13,644 n.a. n.a. n.a. n.a. n.a-
25.8 2.2 1.1 24,000 n.a. n.a. n.a. n.a. n.a. n.a. n.a. 0 n.a. n.a. n.a. n.a. n.a.
4.3 0.9 1.3 14,816 20 44 36 1,625 11 55 0 4,264 n.a. n.a. n.a. n.a. n.a.
2.8 7.7 1.2 3,292 66 20 14 1,950 59 900 27 5,182 n.a. n.a. n.a. 690 13
20.7 1.9 0.9 17,900 50 30 20 1,400 8 1,000 6 1,131 40 40 20 200 18
6.2 0.6 0.3 14,000 69 21 10 810 6 1,670 12 700 39 37 24 n.a. n.a.
8.5 1.6 1.0 11,700 n.a. n.a. n.a. n.a. n.a. n.a. n.a. 4,000 n.a. n.a. n.a. n.a. n.a.
10.3 2.4 1.9 9,140 55 36 9 1,000 11 900 10 7,760 30 40 30 2,300 30
7.8 1.3 1.1 38,449 n.a. n.a. n.a. n.a. n.a. n.a. n.a. 22,504 n.a. n.a. n.a. n.a. n.a.
1.6 0.4 0.7 2,048 70 20 10 450 22 300 15 980 40 30 30 210 21

0.6 0.9 1.7 1,650 27 69 4 637 39 602 36 237 33 33 34 n.a. n.a.

19.1 2.1 1.0 36,630 30 41 29 1,358 4 441 1 12,250 30 5(0 20 n.a. n.a.
- - - 224,124 - - - - - - - 79,612 - - - - -
4.7 1.3 0.9 - 41 35 24 - - - - - 30 46 24 - -

1.3 1.2 0.5 27,056 35 21 44 n.a. n.a. 2,280 8 9,449 20 40 40 n.a. n.a.
317.9 5.3 1.2 1,367 n.a. n.a. n.a. n.a. n.a. n.a. n.a. 0 n.a. n.a. n.a. n.a. n.a.
8.1 11.7 10.2 442 32 48 20 n.a. n.a. n.a. n.a. 1,429 2 43 55 n.a. n.a.
0.9 1.5 2.8 1,351 21 48 31 120 9 n.a. n.a. 8,031 20 30 50 n.a. n.a.
0.8 0.5 0.3 45,291 30 42 28 6,570 15 n.a. n.a. 19,351 30 42 28 n.a. n.a.
- - - 55,887 n.a. n.a. n.a. 8,000 14 8,000 14 106,709 n.a. n.a. n.a. n.a. n.a.
3.0 1.9 1.1 8,964 33 56 11 630 7 750 8 13,867 7 75 18 500 4
2.1 0.9 0.6 9,620 42 37 21 1,730 18 1,953 20 14,777 42 37 22 1,150 8
9.4 1.9 1.6 3,044 22 29 49 890 29 240 8 1,745 8 7 85 n.a. n.a.
7.1 0.6 0.6 2,485 52 10 38 800 32 200 8 990 0 22 78 200 20
3.0 0.9 0.8 3,380 S3 19 28 270 8 790 23 5,020 65 10 25 n,a. n.a.
9.8 1.4 1.2 3,043 7 50 43 243 8 162 5 7,598 11 62 27 n.a. n.a.
8.8 0.5 1.4 606 0 100 0 130 21 130 21 1,867 10 42 48 n.a. n.a.
5.4 1.4 2.0 1,614 50 43 7 n.a. n.a. 282 17 4,404 50 38 12 n.a. n.a.
43.2 2.2 1.9 4,365 10 73 17 222 5 64 1 385 1 56 43 254 66
2.2 0.6 0.3 40,520 85 10 5 4,000 10 2,000 5 2,980 30 50 20 n.a. n.a.
- - - 24,200 n.a. n.a. n.a. n.a. n.a. n.a. n.a. 23,100 n.a. n.a. n.a. n.a. n.a.
10.9 4.0 2.0 3,094 36 54 10 330 11 450 15 5,306 8 68 24 n.a. n.a.
2.8 3.4 2.7 1,695 73 22 5 340 20 470 28 9,546 44 38 18 780 8
2.1 1.5 1.5 7,178 24 24 52 920 13 70 1 20,122 10 30 60 n.a. n.a.
5.6 3.3 1.7 6,445 26 59 15 n.a. n.a. n.a. n.a. 3,368 0 75 25 n.a. n.a.
- - - 251,647 - - - - - - - 260,044 - - - - -
2.7 1.5 0.9 - 44 32 24 - - - - - 24 43 33 - -

- - - 1,025,756 - - - - - - - 792,735 -- - - - -
2.9 0.5 0.9 - 32 42 26 - - 31 3633 - -

43
 APPENDIX

Exploring Cost-Effective
Options for Road
Investment and
Maintenance
This appendix presents a fuller technical explanation these processes on road condition, particularly road
of the findings reported in the main text concerning roughness (one of several such parameters estimated
the economic consequences of different road invest- by the model), which has a significant effect on the
ment and maintenance strategies. It is based on an cost of operating vehicles on the roads.
investigation of the engineering-economicaspects of The model permits the evaluation of several
road deterioration by Bhandari and others (1987). maintenance alternatives for a road link, a group of
roads with similar characteristics, or an entire net-
The Analytical Framework work of paved and unpaved roads. It computes the
aggregate costs of carrying out specifiedmaintenance
The World Bank's Highway Design and Mainte- and construction policies, the associated vehicle op-
nance Standards Study (HDM) and the operational erating costs, and the time streams of total life-cycle
model developed from it (now in its third version, costs discounted at several rates to find net present
HDM-111) provide the framework for the technical values or internal rates of return. These criteria are
analysis of the investment and maintenance options used to determine optimal combinations of road
(Chesherand Harrison 1987;Paterson 1987;Watana- design and maintenance policies. Maintenance poli-
tada, Dhareshwar, and Lima 1987; and Watanatada cies under budget constraints are found by linking the
and others 1987).The model enables investigators to HDM-III model to its companion Expenditure Budget-
evaluate policies, standards, and programs of road ing Model, which employs dynamic programming to
construction and maintenance by applying the results deal with multiperiod resource constraints.
of empirical research on the physical and economic Vehicle operating costs constitute a large share
relationships that underlie the deterioration of roads. (75-95 percent) of the total cost of road transport,
The main variables in the road deterioration cycle except when the traffic volume is extremely low.
and the general relations among them are shown Thus the effect of even a small percentage change in
schematically in box B-1. Most of these effectscan be vehicle operating costs is large relative to the effect
simulated with the model, which contains empirical of changes in construction and maintenance costs.
relationships that have been statistically validated. Before the HDM studies, no empirical data on the
Using data on existing roads and their condition and relationship between vehicle operating costs and the
estimates of future traffic volume and composition, condition of a road's surface (roughness)were avail-
the model predicitsthe deterioration of roads and the able. Investment decisions were guided by imprecise
effects of specified maintenance and improvement and fragmentary estimates of such costs primarily in
policies. It then estimates the cumulative effect of relation to the type of road surface (paved, gravel,

45
 Box B-1. The Road Deterioration Cycle

DETERMINANTS OF DETERIORATION FORMS OF DETERIORATION

* Topographyand subgradesoil Pavedroads Unpavedroads
* Material and thicknessof pavement layers -Potholes Potholes
* Drainage (surfaceand internal) *Ruts * Ruts
* Quality of constructionand maintenance C*
racks Corrugations
* Environment(rain, frost, solar radiation) *Broken edges Erosion
* Traffic(volume,axle loads, and configuration) * Shoulderwear Gravel loss
* Road condition and deformation

MAINTENANCE tructural trengt

* Drainage clearing and vegetation control
* Patching of cracks and potholes
* Shoulder repairs
* Reseals/surface treatment i G
* Pavement strengthening/rehabilitation
* Reconstruction (deferred maintenance)

VEHICLE OPERATING COST

*Speed *Spare parts
* Fuel consumption Maintenance labor
* Tire wear Vehicle utilization and age

Unpaved roads. If they are not maintained, unpaved there is little discernible deterioration and maintenance
roads deteriorate rapidly and reach very high levels of needs are minimal. The next phase, however, is charac-
roughness. The progression of roughness, however, is terized by accelerating deterioration (primarily cracking
essentially linear, and the consequences of poor mainte- and rutting), which leads to increased roughness and, in
nance become apparent immediately. The action of extreme cases, potholing. In the absence of extensive
traffic causes corrugation and rutting and, when com- maintenance operations, this eventually leads to structur-
bined with rainfall, creates deeper ruts, potholes, and al failure. Cracking reduces pavement strength, and poor
even gullies and washouts. Excessive water reduces the drainage and inadequate shoulder maintenance exacer-
road's load-bearing capacity and as a result intensifies bate this effect by allowing rain to penetrate the cracks
the effect of traffic and speeds up the process of deterio- and weaken the underlying layers. Pavement damage
ration. In very dry conditions, the evaporation of water caused by poor drainage is particularly pronounced in
weakens the bond within the surface material, which wet climates and in areas subject to freezing. In one case
then disintegrates under traffic. As the fine binding study, heavy rainfall (150 millimeters a month) reduced
particles turn to dust, the loose gravel left behind needs pavement life by about 30 percent (from fourteen years
periodic replenishment. The wear and tear on roads by to ten) compared with light rainfall (20 millimeters a
traffic is thus strongly influenced by climatic factors. month). Pavement deterioration is critically influenced
Topography and road alignment are also significant by traffic level and loading, with damage increasing
factors, because erosion, vehicle cornering, and hill exponentially with axle load; the damage, however, may
climbing tend to increase gravel displacement. The con- be curtailed by increasing pavement strength in terms of
dition of gravel and earth roads is highly sensitive to the its structural number (SN).In the examples studied (pave-
level and frequency of maintenance. ments with an initial TNof 1.5 to 2.0, and 0.25 to 0.45
million standard axle loads a year), doubling the initial
Paved roads. The progression of deterioration on paved strength postponed the need for an overlay by ten to
roads follows a distinctly nonlinear path. During a long twelve years.
initial phase that lasts up to two-thirds of the life cycle,

46
Table B-1. VehicleOperating Cost Indexes for Paved and Gravel Roads by VehicleClass

Source
Vehicle class de Weille AASHTO
and road type (1966)' (1977)b HDM, 1985'

Passenger car
Paved 100 100 100
Gravel 126-33 108-63 108-26
Light truck
Paved 100 100 100
GraLvel 141-52 109-56 114-38

Heavy truck
Paved 100 100 100
Gravel 140-70 112-58 117-46

Articulated truck
Paved 100 100 100
Gravel 150-70 11341 115-44

Notes: See bibliography for source details.

a. From de Weille (1966), table 11, pp. 28-29.
b. Based on Winfrey (1968), table A-44, p. 727.
c. Based on an application of HDM-1I to Costa Rican data. Steady-state roughness on gravel roads ranged from 5.0 to 9.0 IRI.

earth). Based on improved estimates of vehicle speeds to the circumstances studied in each country, were
(in a free-flowing traffic regime) and operating costs, consistent enough to permit generalizations for wider
as a function of road design characteristics, the HDM application.
research suggests that vehicle operating costs are Thirty-one maintenance policy alternatives were
somewhat less sensitive to changes in road condition tested for paved roads and ten for unpaved roads.
than previously estimated. A comparison of the in- The policies consisted of different maintenance pack-
dexes of vehicle operating costs for paved and gravel ages with specifications of the deterioration levels at
roads shows significant differences between the HDM which they would be applied. The options for paved
results and estimates from two earlier standard works roads ranged from pothole patching, to bituminous
(see table B-1). resealing of the entire surface and more costly over-
Improvements in vehicle design over the past twen- lays of asphalt concrete, to major rehabilitation and
ty years may in part account for the relatively lower reconstruction of the base and the surface (see the
estimates obtained from the HDM relationships. Nev- glossary at the front of this book for definitions of
ertheless, certain road improvements may have the various types of road improvement and mainte-
smaller benefits (in vehicle operating costs saved) nance works). A null case-comprising only those
than previously estimated, and the inadequacy of routine maintenance activities (drainage clearing,
past data may have contributed to the premature minimal vegetation control, and shoulder repair)
paving of lightly trafficked roads, particularly in included in all the other alternatives-served as a
Africa, and to the underdesign of heavily trafficked benchmark against which the costs of the alternatives
national roads in India, Indonesia, Nigeria, and Pa- were measured and compared. The alternatives are
kistan. summarized in tables B-2 and B-3, where each is
identified by a code (such as AL18). These codes are
Case Studies used throughout this appendix in discussing the re-
sults. Under policy AL18, for example, all potholes
HDM-iII and the Expenditure Budgeting Model were would be patched every year, a surface treatment
employed in a series of case studies to investigate would be applied whenever 25 percent of the area
optimal investment and maintenance policies under was visibly damaged, and a 40-millimeter overlay
a variety of conditions. Road types and conditions, would be applied whenever roughness reached 5.0 on
traffic, climate, and unit costs in the studies corre- the International Roughness Index (IRI).
sponded to those observed in three countries-Chile, The roads in each country's network were grouped
Costa Rica, and Mali. The results, although specific into broadly homogeneous classes according to sur-

47
Table B-2. Maintenance Policy Alternatives for Paved Roads

Patching Reconstruction
(percentage
of potholes Asphalt concrete overlay Bituminous
patched surface Asphalt
Alternative annually) Resealinga 40-millimeter 80-millimeter treatment concrete

AL0O (null) 0
AL01 50
AL02 100
AL03 0 At 8.5 [RI

AL04 50 At 8.5 [RI

ALOS 100 At 8.5 IRI
AL06 100 At75 % damage
AL07 100 AtS50%damage
AL08 100 At 25% damage
AL09 100 At 75% damage At 8.5 IRI
ALIO 100 At 50% damage At 8.5 IRI
AL11 100 At 25% damage At 8.5 IRI

AL12 100 At 5.0 IRI

AL13 100 At 4.2 IRI
AL14 100 At 3.5 IRI
AL15 100 At 5.0 IRI Immediateb

AL16 100 At 4.2 [RI Immediateb

AL17 100 At 3.5 [RI Immediateb
AL18 100 At 25% damage At 5.0 [RI
AL19 100 At 50% damage At 3.5 [RI
AL20 100 At S.0 IRI Immediateb
AL21 100 At 4.2 IRI Immediateb
AL22 100 At 3.5 IRI Immediateb
AL23 0 At 8.5 [RI
AL24 50 At 8.5 [RI
AL25 100 At 8.5 IRI
AL26 100 At 75% damage At 8.5 IRI
AL27 100 At 50% damage At 8.5 [RI

AL28 100 At 25% damage At 8.5 IRI

AL29 100 At 5.0 iRi Immediateb
AL30 100 At 4.2 [RI Immediate'
AL31 100 At 3.5 [RI Immediateb

Notes: Basic routine maintenance such as drainage clearing, minimal vegetation control, and repair of shoulders and drains is included in each
alternative.
a. Resealing with bituminous surface treatment whenever the area of pavement surface with specific signs of distress and visible defects reaches
the percentage shown.
b. During the first year of the period studied, followed by a maintenance strategy involving patching, reseals, and overlays at the intervals
specified.

face type, condition in the initial year of the study, Expenditure Budgeting Model, the results for differ-
and traffic volume. For each class, road deterioration ent road classes were combined to find optimal
and maintenance activities over a thirty-year period network strategies, costs, and benefits under condi-
were simulated under various maintenance alterna- tions involving no budget constraints and also under
tives. Maintenance costs and vehicle operating costs varying levels and time periods of overall budget
were computed, discounted to the initial year, and constraints. The detailed results from the case studies
subtracted from costs for the null case. The results were used to explore cost-effective road investment
gave the net present value for each alternative relative and maintenance options. The main findings have
to the null. From these results the best strategy for been grouped into the following topics for discus-
different levels of funding and discount rates was sion:
determined for each road class. With the aid of the * Selection of cost-effective maintenance policies

48
Table B-3. Maintenance Policy Alternatives for Unpaved Roads

Spot Gravel
Alternative Blading frequency regraveling, resurfacing'

AL0O (null) None No No

AL01 Once a year Yes No
AL02 Every 8,000 vehicle passes Yes No
AL03 Every 6,000 vehicle passes Yes No
AL04 Every 4,000 vehicle passes Yes No
AL05 Every 2,000 vehicle passes Yes No
AL06 Once a year Yes Yes
AL07 Every 8,000 vehicle passes Yes Yes
AL08 Every 6,000 vehicle passes Yes Yes
AL09 Every 4,000 vehicle passes Yes Yes
AL10 Every 2,000 vehicle passes Yes Yes

Notes: Basic routine maintenance such as drainage clearing, minimal vegetation control, and repair of shoulders and drains is included in each
alternative.
a. Replacement of 30 percent of total gravel loss each year.
b. Resurfacing with a gravel layer 150 millimeters thick whenever thickness of existing surface falls below 50 millimeters.

* Optimization of maintenance expenditures under costs. In most cases this yields a positive saving, but
budget constraints alternatives AL15, AL16, and AL17 (involving the
* Criteria for choosing pavement strength immediate application of a thick overlay) and AL20,
* Economic traffic thresholds for paving gravel AL21, and AL22 (involving immediate reconstruc-
roads tion) cost more than the benefits they yield when
i The impact of overloading. discounted at 12 percent.
Net present value, again discounted at 12 percent,
Selection of Cost-Effective Maintenance Policies is maximized by strategy AL12 (patching all pot-
holes annually and applying a 40-millimeter overlay
The choice and timing of maintenance operations on whenever pavement roughness exceeds 5.0 IRI). But
paved and unpaved roads are strongly affected by the several other alternatives, with very different combi-
differences in their deterioration characteristics. With nations of vehicle operating costs and road agency
unpaved roads, whose deterioration is both linear expenditures, are almost as good by this measure.
and rapid, special attention to routine maintenance This finding is important because it widens the room
is required, particularly with regard to the frequency for maneuver when agencies are subject to budget
of blading. With lpaved roads, the nonlinear path of constraints. For example, the curve in graph B, which
their deterioration permits more options as to the relates agency expenditures to vehicle operating costs
choice and timing of maintenance. for selected alternatives, shows that AL05 (pothole
To identify cost-effective maintenance policies for patching until roughness reaches 8.5 IRI, then recon-
paved roads, particularly under conditions involving struction) entails about half the road agency expen-
budget constraints, the tradeoff between agency and diture of AL12, while vehicle operating costs are
user costs was examined. Figure B-1 shows the results higher by about $1.10 for each agency dollar saved-
of an analysis of twenty-two maintenance alterna- a net loss of only $0.10 per $1.00 of reduction in
tives for a low-volume paved road in good to fair agency expenditure. If funds are limited, this may be
condition in Mali, with an average daily traffic of 400 an attractive option.
vehicles. (Only twenty-two of the thirty-one alterna- The net present value of total transport costs for
tives specified in table B-2 are depicted because different levels of agency expenditures is illustrated
results for some pairs of alternatives were identical.) in figure B-2 for four cases. The most efficient mainte-
The top chart (A) shows the present value, discount- nance alternatives lie on the positively sloped seg-
ed at 12 percent, of the net cost saving of each ment of the outer boundary-the efficiency frontier.
alternative relative to the null case. The net cost Alternatives represented by points inside the frontier
saving is determined by subtracting the increase in are always inferior to a combination of maintenance
maintenance and construction costs incurred by the options lying on the frontier. For example, in case A
road agency from the difference in vehicle operating (Mali), AL09 has a net present value of $12,000 a

49
Figure B-1. Analysis of Maintenance Alternatives for a Low-Volume Paved Road in Good to Fair
Condition (Mali)

A. Net present value of maintenance alternatives

30 -

20 -
E
0_
10 E
LOAL2AL03 ALO4 AL05AL6 AL07 ALOS AL19AI
AI L2 AL13 AL14 A L1SL9
0 ~ '-t-~---~-' '.._-~---~' ~ ALIS AL16 AL17 ' AL20 AL21 AL22
[~ -30 -0 uD -of

~-40 X
680
0

co . .c
. ( s
030
-40
~ ~
(tbousands ~
prklontr ~ ~ ~ ~ C of doZr
a. With bituminous surfacetreatment.

B. Vehicle operating costs versus road agency costs

(thousands of dotlarsperkilometer)
uo \90 (pthn n eosrcin
680
ALOO(null alternative)

670 AL02 (patching only)

n~660

AL05(patching and reconstruction)

650

AL12 (overlays at 5.0 IRI)

0. 640 - AL13 (overlays at 4.2 ]RI)

630
0 10 20 30 40 50
Present value of agency expenditures

Note. Seetablc B-2 for specifications

of the maintenance
altemnatives.

50
Figure B-2. Net Present Value versus Road Agency Expenditures for Selected Maintenance Strategies
(thousands of dollars per kilometer)

A. Low-volume (ADT = 400) paved road in good to fair B. Medium-volume (ADT = 700) paved road in poor
condition (Mali) condition (Chile)
20 400 AL22
350 -AL20 AL16
16 ALO AAL5L2 30 _L2 _ 1
-- ------ AL13 > 300 - AL03 AL12 *
9AL18~~~~~~~~~~AI
> AL8250 ~~LO > - AL18
12 * *AL23 ALLS AL23
ALO9 ~~~~~~~~ ~200-
^. 8 *AL06 : 150 -
4 ALOI 00- AL02
50

0 5 10 15 20 25 30 0 20 40 60 80
Present value of agency costs Present value of agency costs

C. Medium-volume (ADT =800) paved road in fair condition D. High-volume (ADT =1,500) paved road in poor condition
(Costa Rica) (Costa Rica)

450 AL13600
400 6AL4 AL13 AL16 AL17
500 A1
, 350 ALO8 L18 * * AL23
, 300 -AL3 > 400 AL40 AL03
250-
~
ALO2 ~ ~ ~ ~ ~~~ ~300
u 200 -
t 150 - 200
z z
100 0
50 LOOO

0 10 20 30 40 50 0 10 20 30 40 50 60
Present value of agency costs Present value of agency costs

Note: See table B-2 for specifications of the maintenance alternatives.

kilometer at a present value of agency cost of $15,800 The curve also shows the optimal order of retract-
a kilometer. A higher net present value can be ob- ing when agency resources are cut. The lower-order
tained at the same agencycost by using AL05on part options are selected by successivelyreducing agency
of the road group and AL12 on the rest; the optimum expenditures in ascending order of their marginal
maintenance strategy then lies on the line connecting contribution to net present value. Where traffic is
AL05 and AL12 on the graph. In the absence of heavier, and especially where roads are initially in
budget constraints, the alternative that maximizes poor condition (cases B and D), the economic loss per
net present value should be selected. Agency expen- dollar saved by the agency is considerably higher, and
ditures beyond this point (AL12) yield benefits that of course successive reductions in expenditure have
are less than the increased costs, as reflected by increasingly costly consequences. The steepest slope
declining net present value. of the efficiency frontier, indicating the severest pen-

51
alty, is reached when the only expenditure avail- treatment drops to 25 percent or less of that for an
able for cutting (always excepting minimal routine overlay.
maintenance) is pavement patching. On unpaved roads the primary maintenance-relat-
A summary of the maintenance options recom- ed determinant of roughness-and so of the cost of
mended for paved roads, based on generalizations operating vehicles-is the frequency of surface blad-
drawn from the case studies, is presented in table B-4. ing. Simulation and cost analysis of the effects of the
These recommendations reflect interactions between various maintenance options (see table B-3) show
traffic levels and road conditions and the mainte- that blading costs and vehicle operating costs are
nance interventions needed to reduce the total cost of closely balanced over a wide range of blading fre-
road transport. In individual cases the level and quencies. The optimal frequency is in the range of
frequency of optimal maintenance actions may vary one blading every 4,000 to 8,000 vehicle passes,
significantly from those recommended because of which is consistent with what is generally accepted
variations in construction and maintenance prac- as good practice (for example, one blading a month
tices, traffic loading characteristics, environmental at 150-200 vehicles a day). Incremental increases in
conditions, or factor costs. For example, the results blading frequency within this range result in mar-
of the case studies suggest that surface treatments ginal reductions in user costs that are almost com-
become a viable economic alternative to asphalt pletely offset by corresponding increases in agency
overlays when the cost per kilometer for surface cost. Figure B-3 shows plots of net present value

Table B4. Recommended Maintenance Options for Minimizing Transport Costs on Paved Roads

Averagedaily
traffic(number Initialroad
of vehicles) condition, Recommendedmaintenance
Fewer than 200 Good to poor Pothole patching only until roughness level becomes very high (8.5 IRI) and the
traffic volume has risen sufficiently (ADT> 200) to warrant pavement
reconstruction.

200 to 500 Good to fair Pothole patching and overlays (for example, 40-millimeter asphalt concrete
whenever roughness reaches 4.2 to 5.0 IRI). Under budget constraints, surface
treatments may be substituted for overlays.
Poor Immediate reconstruction of the pavement.b

500 to 1,000 Good Pothole patching plus periodic overlays (40-millimeter asphalt concrete at 3.5 to
4.2 IRI).
Fair Same as above but with an initially thicker overlay (80-millimeter asphalt
concrete) in cases of weak pavement.
Poor Immediate reconstruction of the pavement.6

1,000 to 2,000 Good to fair Pothole patching plus periodic overlays (40-millimeter asphalt concrete at 3.5
IRI). Where the existing pavement is weak, the initial overlay should be thicker
(80-millimeter). Periodic surface dressings are economical on strong pavements
in good condition.
Poor Immediate reconstruction of the pavement.)

More than 2,000 Good Pothole patching plus periodic surface dressings (when 25 percent of the area is
damaged) in addition to periodic overlays (40-millimeter at 3.5 IRI).
Fair Pothole patching plus periodic overlays (40-millimeter asphalt concrete at 3.5
IRI). Where the existing pavement is weak, the initial overlay should be thicker
(80-millimeter).
Poor Immediate reconstruction of the pavement.b

a. Good = less than 3.5 IRI;fair = 3.5 to 5.8 [RI;poor = greaterthan 5.8 IRI.
b. Some provisionfor patchingand emergencymaintenanceshould be made to keep the road serviceableduring the reconstructionperiod.

52
 could prolong the time between bladings with me-
Figure B-3. Variation of Net Present Value with chanical graders. For unpaved roads with very low
Blading Frequency(Costa Rica) traffic volumes, the increased use of labor for spot
repairs could offer a viable alternative to mechanical
______________________________________means of maintenance (Coukis and others 1983).The
lower maintenance costs would offset the somewhat
higher vehicleoperating costs.
Gravel surface ADT 300
Optimizationof MaintenanceExpenditures
c 400 - E surface_under Budget Constraints

When the maintenance budget for a network is less

than that required for the overall optimum, alloca-
n 300 =200 tions to specific road classes (defined on the basis of
o -- - ADT design, traffic, and road condition and, where appro-
- <-_ ~ - priate, further stratified by regional, climatic, or
other relevant factors) must be reduced from their
200 optimal levels. But maintenance expenditures should
not be reduced uniformly across all road classes. The
ADT= 100 loss in net present value can be minimized by first
__________________________ Lreducing allocations to those roads for which the
efficiency frontier is least steep. In general, candidates
zADT = 50 for cutting are maintenance operations with high unit
z costs on roads with low volumes of traffic and good
ADT=25 existing surfaces. After the optimal activities have
0 , I I I I 1 been completely replaced by the next-best options in
0 2 6 4 8 10 12 14 a road class, the penalty for further cutbacks will
Vehicle passes between bladings be proportionately greater-the frontier will be
(thousands) steeper-and it may be economical to reduce alloca-
tions to some other road classes as well. Roads with
very high volumes of traffic and poor surface condi-
tions suffer the greatest loss in benefits for each dollar
of reduction in maintenance outlay, and their alloca-
versus blading frequency obtained from the Costa tions should be reduced last.
Rican study. Even at low traffic levels (twenty-five The effects of reductions in the maintenance bud-
vehicles a day), the economic returns on blading are get on the choice of alternatives and on the resulting
substantial. Blading once a year appears to be an benefits were examined for the road networks of
acceptable minimum threshold. Although blading Chile, Costa Rica, and Mali. In Costa Rica the
once after every 4,000 to 8,000 vehicle passes would maximization of total net benefits would require
be an optimal policy, less frequent blading does not spending an average of $12.5 million a year on
occasion serious economic loss if the road is regrav- maintenance for the first ten years, with $38 million
eled at appropriate intervals. Local conditions will needed in the first year alone to rehabilitate the paved
strongly influence the maintenance options for un- roads in poor condition. Figure B-4 shows the max-
paved roads. Extreme combinations of soils and imum net present value of benefits (discounted at 12
climate (sands in arid climates, heavy clays in wet percent a year) for the Costa Rican road network
climates) may justify surfacing earth roads with grav- under different levels of average annual maintenance
el at an average daily traffic flow of fewer than fifty expenditures, optimally allocated among different
vehicles to ensure accessibility. classes of road. If the budget is raised from the
Where capital-especially foreign exchange-is presently planned level of $6 million a year to the
scarce and unskilled labor is available at competitive optimal level of $12.5 million a year over twenty-five
rates (currently about five dollars or less a day), using years (or by $51 million in present value), the attain-
labor-based work methods to perform spot repairs on able net present value (with the best use of the funds
gravel roads and simple drags attached to agricultur- in both cases) increases by $200 million, from $635
al tractors or trucks to even out surface corrugations million to $835 million. The optimal program, even

53
 reduced for any reason, the losses would be large for
Figure B-4. Variationof Net Present Value of the this highly trafficked network.
EntireRoad Network as a Function of Average In Mali, by contrast, where 84 percent of the paved
Annual Maintenance Expenditures (Costa Rica) roads carry fewer than 200 vehicles a day, only
(millionsof dollars) patching and basic routine maintenance is economi-
cally justified on most roads. The most economical
option is to keep only about 1 percent of the paved
900 network in good condition and maintain the rest
of the network at considerably reduced standards,
800So mainly with patching and routine maintenance on
paved roads and minimal blading on unpaved roads.
700 - Even so, about $9 million is required to clear Mali's
backlog of economically warranted rehabilitation
_________ projects for high-volume paved roads currently in
600 / 0 poor condition. And the average annual expenditure
required to maintain the combined network of paved
> 500 and unpaved roads is estimated at $6.2 million a
year-about twice the current expenditure. If expen-
c. 400 _ / I ditures were raised to $6.2 million ($24.3 million in
present value over twenty-five years), the present
z Projectedlevel value of the benefits is estimated at more than $46
300 j of expenditure million.
Maintenance activities are often deferred during
200 I periods of austerity. For unpaved roads, as long as
basic routine maintenance is carried out regularly,
the primary effect of deferring blading and, to a lesser
100 extent, regraveling is to increase vehicle operating
costs during the deferral period. The effect on subse-
0 2 4 6 8 10 12 14 16 quent
road isroad restoration
allowed costssois bad
to become large,
not that it unless
has to the
be
Average
annual maintenanceexpenditure, reconstructed,generallyon a new alignment.For
paved roads, however, both effectscan be important:
a. Ten-yearaverage. vehicleoperatingcosts increaseduring the deferral
period, and the cost of pavement rehabilitation can
vary substantially, depending on the stage in the
deterioration process at which deferral occurs. On
newly constructed or rehabilitated pavements with
with unlimited funds, would not keep all roads in light axle loading, the effect of deferring maintenance
good condition. In Costa Rica it would provide for a (other than basic activities such as drainage) for one
high level of maintenance to keep two-thirds of the to five years is negligible. Once pavement condition
paved roads in good condition, with the other third becomes fair or poor, the impact is large.
to be maintained at lower standards.
For Chile the best maintenance strategy for the Criteriafor ChoosingPavementStrength
road network is similar to that for Costa Rica: two-
thirds of the paved roads to be kept in good condition When a new pavement is constructed or an exist-
and the rest maintained at lower standards. The ing one replaced or overlayed, the choice of design
optimal program involvesthe immediate expenditure strength should take account of the reliability of
of large sums for the rehabilitation of paved roads future maintenance. Low probabilities of adequate
and assigns high priority to such rehabilitation even maintenance and timely strengthening in the future
if the budget has to be cut back. That is the course favor building a strong pavement initially, since
Chile actually adopted, and the planned budget for stronger pavements enjoy a longer grace period dur-
maintenance of all roads-$140 million a year during ing which maintenance needs are minimal. A normal
1984-91-is approximately the amount required for full-strength pavement is defined as one designed to
the optimum derived in this study. If the budget were carry a specific number of cumulative equivalent

54
 standard axle loads (ESAL) for a given level of pave-
Figure B-5. Influenceof Maintenance Policies on ment serviceability. To compensate for inadequate
the Selection of Initial Pavement Strength maintenance, a pavement strength higher than that
given by normal designs may be warranted in certain
A. Light traffic (ADT= 500), lightly loaded (0.04 million ESAL situations. High probabilities of good maintenance
per yearper lane) will favor time-staging-that is, economizingon to-
day's pavement and strengthening the road subse-
> 700 quently as needed. At the network level, time-staging
. E _ SN = 35 is effective only if the condition of pavements is
~= 680
oid N 35 regularly monitored and evaluated. Without an ap-
propriate pavement management system, it is dif-
660 i
:; ficult to predict the critical points in the pavement
o_ \< __ SN = 2.0 life when a major maintenance intervention may be
0 _ __-_ needed to prevent premature structural failure.
To justify the time-staging of road construction, a
620 minimum probability of adequate future mainte-
600 l l l l l l l l l l nance is required.To estimatethis thresholdprob-
ALOI AL03 AL05 AL07 AL09 AL11 AL13 ability, life-cyclecosts for pavements, with an initial
Maintenance alternative structural number (SiN)of 2.0, 3.5, and 5.0, were
estimated for a range of traffic volumes and axle
B. Medium traffic (ADT= 1,000), lightly loaded (0.08 million loadings in Costa Rica (light) and Mali (heavy).
ESALper year per lane) Three of the cases are illustrated in figure B-5. Each
1.34 case shows life-cyclecosts (discounted at 12 percent)
for the first fourteen of the thirty-one maintenance
1.32
alternatives for paved roads specifiedin table B-2. In
= \ " case A, with an average daily traffic (ADT) of only 500
E; 1.30 \ l\ r vehiclesand lightaxle loading,it wouldsufficeto use
o0 1 28 _ \,I / s, SN = 3.5 a normal design (SN= 2.0) commensurate with the
_ 1.28 z _ -/estimated ESAL, since the life-cycle costs are consis-
tently higher for a higher-strength compensating de-
UO . SN = 2.0 \\ sign (SYN= 3.5) under all maintenance assumptions.
i3
1.24 _ s But with heavier traffic a different conclusion
E x, emerges, as shown by case C (ADT = 2,500 and heavy
1.22 1 I 1
I, I I, E axle loading). In this case a normal full-strength
ALOIAL03 AL05 AL07 AL09 ALII AL13 zpavement (SN= 5.0) will have a lower life-cyclecost,
Maintenance alternative under almost any assumption about future mainte-
nance, than a pavement with a lower initial strength
C. Heavytraffic(ADTL2,500), heavilyloaded(1.21million (SN = 3.5), whichrepresentsthe time-stagingoption.
ESALper year per lane)
At or above this combination of traffic and axle
, 4.5 loadings it would not pay to consider the time-
staging option.
'AO 4.4
Between these two limits, the decision for or
E <:. 4.3 _ A / against the time-stagingof construction is reached by
r \ balancing
/\ the loss that the lack of future maintenance
5 o 4.2 r \J \ could occasion against the cost saving from time-
staging that would be realized if future maintenance
4.1 - 50 --- (including strengthening)wereperformedas desired.
i- = In this way a threshold probability of performing
4.0 l l l l good maintenancein the future is obtained;above
ALOI AL03 AL05 AL07 AL09 ALII AL13 this threshold the time-staging option could be justi-
Maintenance alternative fied. For case B in figure B-5, for example, with an
Notes: Seetable B-2for specifications
of the maintenancealternatives. ADT of 1,000vehiclesand light axle loads, the thresh-
a. Presentvaluediscountedat 12percent a year. old probability was estimated at 33 percent as fol-
lows:

55
 If expedient to forgo maintenance until it becomes
x = probability of optimal maintenance necessary to reconstruct the pavement. AL03 in case
(AL14) B codifies such a maintenance strategy, an option
1 - x = probability of nil maintenance (ALOO) that would also reduce life-cyclecosts significantlyif
D.'= difference in life-cycle costs for initial there is a fair expectation that funds for reconstruc-
pavement design strength of SN = 3.5 and tion will be forthcoming when needed. In this case a
SN = 2.0, under AL14 less than full-strength design could be considered for
D.d = difference in life-cycle costs for initial initial construction if the probability of obtaining
pavement design strength of SN = 2.0 and future funds for reconstruction is 45 percent or better.
SN = 3.5, under ALOO Table B-5 illustrates how the decision may vary in
Then, from the inequality: a wider array of cases. It maps out combinations of
(D.pt)(x) > (D0,) (1- x) discount rate, daily traffic, and axle loadings in which
or (1.27- 1.23) (x) > (1.33-1.31) (1 -x) the decision for or against time-staging is inde-
Therefore, x > 0.33. pendent of the degree of uncertainty about future
maintenance. In a middle area, however, the reliabil-
In this case the time-staging of pavement construc- ity of future maintenance matters. Time-staging is
tion should be considered only if there is a one-third thus generally to be preferred if the probability of
or better chance that good maintenance will be adequate maintenance in the future exceeds 30 or 75
performed in the future. percent, depending on trafficvolumes and axle loads.
Under funding arrangements that favor construc- At higher discount rates (in this case 24 percent),
tion over maintenance or in situations in which time-staging tends to be the preferred choice even
external aid agenciesare willing to finance construc- with high trafficvolumes, heavy axle loadings, and a
tion but not maintenance expenditures, it is often relatively low probability of future maintenance.

Table B-5. Criteriafor SelectingTime-StagedPavementDesign

Average Minimum probability of adequate maintenance

daily required for time-staging optionb
traffic
(number of Axle At a discount At a discount At a discount
vehicles) loading' rate of 6% rate of 12% rate of 24%

300 Light * * H
Heavy 30% * H

500 Light 30% * H

Heavy n.a. 75% H

1,000 Light 95% 30% U

Heavy n.a. 80% H

2,000 Light 95% 80% 15%

Heavy 0 90% 15%

2,500 Light * 95% 50%

Heavy 0 0 60%

Notes: The economic costs for paving (in thousands of dollars per kilometer in 1984prices), assuming all earthworks and structures already in
place, were estimated as:
SN Costa Rica Mali
2.0 19.6 27.2
3.5 68.5 59.3
5.0 118.8 129.4
a. Light axle loading is representative of conditions in Costa Rica (0.05-0.10 million ESAL per year per lane); heavy axle loading is representative
of conditions in Mali (1.2 million ESAL per year per lane).
b. Time-staging of pavement construction is conditional on a minimum probability of adequate maintenance in the future. If the probability
of adequate maintenance is below this level, a normal full-strength design should be used.
* Compensatory design not applicable; low- to medium-strength pavements (SN< 3.5 based on normal design) adequate for this level of traffic
volume and loading.
* Time-staging not applicable; high-strength pavements (N>4.0 based on normal full-strength design) appropriate.
n.a. Time-staging option not applicable; use normal-strength design.

56
Where capital is cheap or a greater weight is attached
to long-term benefits (an intergenerational issue), Figure B-6. Break-even Traffic Volumes for
stronger initial pavements based on normal design Paving a Gravel Road under Good and Poor
generally dominate (say, at a discount rate of 6 Maintenance
percent).

Economic Traffic Thresholds for Paving Gravel E 2 200

Roads .
* 160
Even with the best maintenance practice, vehicle 120
operating costs on gravel roads are between 10 and > m
30 percent higher than on paved roads. In addition, o 8
the present value of the cost of routine mainte- ,
nance and resurfacing is between five and eight > 0 40 - 80 310
times greater for a well-maintained, high-volume
gravel road than for a newly built paved road. Paving Q G aitnc
is therefore indicated when the expected savings in o
-40 i I I
vehicle operating and maintenance costs (relative to Z 40 80 120 160 200 240 280 320 360
a well-maintained gravel road) exceed the present Initialaveragedailytraffic (numberof vehicles)
value of the paving cost.
These cost tradeoffs seem to cover a wider range
of trafficvolumes than had commonly been assumed.
Applications of HDM-III in Costa Rica and Mali
indicate that the break-eventraffic volume for paving It was assumed that the same level of basic routine
may vary from less than 100 vehicles a day to more maintenance (vegetation control, drainage clearing,
than 400, depending on the costs of paving, the and so on) would be performed in all cases. With
discount rate, the rate of growth of traffic, and the good maintenance, the economic threshold for pav-
anticipated quality of future maintenance. ing a gravel road occurs at a traffic level of 310
Paving thresholds are also sensitive to variations vehiclesa day. With poor maintenance, however, the
in assumptions about the quality of maintenance. break-even traffic volume for paving is lowered to 80
Where experience shows that the probability of ade- vehiclesa day. Inadequate maintenance increases the
quate future maintenance is low, the trafficthreshold economic and political pressures on a road agency to
for paving is lowered. And if there are no budget make larger pLiblic investments in paving gravel
constraints, a case can be made for an all-weath- roads which, had they been well maintained, would
er paved road. Such roads may not require major have continued to provide a reasonable level of
maintenance work for as long as seven to ten years. service.
With regard to total life-cyclecosts, however, more Paving gravel roads in arid zones is sometimes
frequent regraveling and grading operations carried suggested as a way of alleviating the discomfort and
out efficiently are likely to prove more economical inconvenience of travel on dusty roads, but it is
than paving, particularly if the future availability of difficult to quantity these benefits. Common observa-
maintenance funding is uncertain. tion suggests that vehicle speeds and passing oppor-
The effect of quality of maintenance on paving tunities on such roads are severely restricted by
thresholds is illustrated in figure B-6,which compares reducedvisibility.The result is a traffic hazard similar
two maintenance regimes in Mali. to fog and congestion and akin to that caused by
* Good maintenance. For the paved alternative: 100 heavily loaded slow-moving vehicles on narrow
percent patching and resealing every seven years. roads. Not enough empirical information is available
For the existing gravel road: grading every 4,000 on traffic flow and vehicle operating characteristics
vehicle passes and regraveling whenever gravel in a dusty environment to evaluate the benefits of
thickness falls below 6 centimeters. paving roads in arid or desert areas.
* Poor maintenance. For the paved alternative: no Lower paving thresholds may also be indicated for
patching or resealing, but reconstruction whenever roads located in river deltas (in lower Bangladesh,for
roughness exceeds 12.0 IRI. For the existing gravel example), old lake beds, sandy deserts, or low coastal
road: infrequent grading (about once a year) and areas, because of the scarcity of gravel deposits and
no regraveling. other sources of aggregate. In mountainous terrain,

57
where maintenance costs may be excessiveowing to
the erosion caused by rapid runoff and the frequent Figure B-7. Effect of TrafficLoading on Pavement
need to replace gravel, it may make economic sense Performance
to protect the road with a bituminous surface. A
technical alternative to using a gravel surface in these
situations is to stabilize the existing soils with small A. Heavyloading(cumulative
ESAL=14.0 million)
amounts of a suitable binder (bitumen, cement, lime, x 12
or fly ash) and then apply a light bituminous seal, _
where necessary, to protect the stabilized material Vi 10
from the elements and traffic. 8

The Impactof Overloading 0i 6

4
The onerous impact of overloading on pavement .......
performance is reflected in the progression of rough- E 2
ness over time, given different maintenance policies 0 ' '
and loading conditions on otherwise similar pave- 0 4 8 12 16 20 24
ments (see figure B-7). Case A is typical of Mali, Time (years)
where trucks with heavy axle loads constitute 33
percent of the traffic and the cumulative ESALper lane
over twenty-five years is estimated at 14 million. Case B. Light loading (cumulative EsAL 2.5 million)
B is typical of Costa Rica, where trucks with rela-
tively light axle loads constitute 25 percent of the n 12 -
traffic and the cumulative ESALper lane over twenty- m 10 _
five years is estimated at 2.5 million. The actual axle r_ 8-
loads, on average, are about 60 percent heavier in 6
case A, partly because of the higher axle load limits X 6-
(thirteen-ton single axle) in Mali and partly because C . 4
of overloading. Heavy loading significantly acceler- 2...... ...
ates the rate of deterioration, and in this comparison E
the need for overlays in case A is advanced by some 0 4 8 12 16 20 24
six years. Even with optimal maintenance, the aver- Time (years)
age life-cycle roughness remains higher with heavier
axle loads. The results for case A are typical of what No maintenance Patching
only Resealing
has happened to pavements designed on the basis of - Reconstruction Overlays
an unrealistic assumption that axle load limits would
be enforced. Note: Data are basedon the followingassumptions:SN 3.0 and ADT=800.
Although empirical evidence shows that a signifi-
cant part of pavement damage increases with the
fourth power of the axle load, the regulation of axle
loads has proved exceedingly difficult. Many road
authorities have therefore built stronger and more
expensive pavements, based on actual pavement increased beyond the prevailing eight- to ten-ton
loading, than would have been necessary with ade- single (and thirteen- to sixteen-ton tandem) axle
quate control of axle loads. Expected changes in road loads; the economic gain from lower unit transport
transport technology resulting from road improve- costs outweighs the loss in increased road damage. In
ments (such as the introduction of combination vehi- the short term, however, the existing roads and
cles in place of single-unit trucks) should be taken bridges are not strong enough to carry heavier loads.
into consideration when pavements are designed. Any increase in axle load limits should therefore be
Research on the economic optimum for axle loads accompanied by a comprehensive plan for strength-
suggests that limits in the longer term should be ening bridges and pavements.

58
 Bibliography

Background papers commissioned for this study are Chesher, Andrew, and Robert Harrison. 1987. Vehi-
indicated by an asterisk. The word "processed" de- cle Operating Costs: Evidence from Developing
scribes works that are reproduced from typescript by Countries. Bailtimore, Md.: Johns Hopkins Univer-
mimeograph, xerography, or similar means. Such sity Press.
works may not be cataloged or commonly available Choate, Pat. 1983. Bad Roads: The Hidden Costs of
through libraries or may be subject to restricted Neglect. Riverdale, Md.: National Asphalt Asso-
circulation. ciation.
Cooper, Lauren. 1984. The Twinning of Institutions:
AASHTO. American Association of State Highway and Its Use as a Technical Assistance Delivery System.
Transportation Officials. 1977. A Manual on User World Bank Technical Paper no. 23, Washington,
Benefit Analysis of Highway and Bus Transit Im- D.C.
provements. Washington, D.C. Coukis, Basil, and others. 1983. Labor-Based Con-
Baker, Michael. 1984. Rebuilding America's Infra- struction Programs: A Practical Guide for Planning
structure: Agenda for the 1980's. Durham, N.C.: and Management. New York: Oxford University
Duke University Press. Press.
Benor, Daniel, James Q. Harrison, and Michael Bax- *Cox, Brian E. 1987. "Evaluation of Incentives for
ter. 1984. Agricultural Extension: The Training Efficiency in Road Maintenance Organization:
and Visit System. Washington, D.C.: World Bank. The U.K. Experience." Transportation Depart-
Bhandari, Anil, and Kumares Sinha. 1979. "Optimal ment Discussion Paper, Transportation Issues Se-
Timing for Paving Low Volume Gravel Roads." ries, Report no. TRP4, World Bank, Washington,
In Transportation Research Board. Transporta- D.C. Processed.
tion Research Record 702. Washington, D.C. de Weille, Jan. 1966. Quantification of Road User
"Bhandari, Anil, and others. 1987. "Road Deteriora- Savings. World Bank Staff Occasional Paper no. 2,
tion in Developing Countries: Technical Options Washington, D.C.
for Road Maintenance and Economic Conse- Faiz, Asif, and Esra Bennathan. 1988. "Road Deteri-
quences." In Transportation Research Board. oration in Developing Countries: The Financial
Transportation Research Record 1128. Washing- and Institutional Challenge." Paper no. 265, 14th
ton, D.C. Conference of the Australian Road Research
Buchanan, James A. 1963. "The Economics of Ear- Board, Canberra, August 28-September 2. Pro-
marked Taxes." Journal of Political Economy (Oc- cessed.
tober). Faiz, Asif, and Per Fossberg. 1987. "Truck Weight

59
 Control Experience in Developing Countries." In International Road Federation. 1985. World Road
Transportation Research Board. Supplement to Statistics, 1980-84. Washington, D.C.
Transportation Research Record 1106 (Proceed- International Road Transport Union. 1985. World
ings of the Fourth International Conference on Transport Data. Geneva.
Low-Volume Roads, Ithaca, New York, August). *Johansen, Frida. 1984. "Deterioration of the Road
Washington, D.C. Networks-an Emerging Crisis." Washington,
*Faiz, Asif, Clell Harral, and Frida Johansen. 1987. D.C.: Transportation Department, World Bank.
"Road Deterioration in Developing Countries: Mason, Melody. 1985. "Road Maintenance Survey
The State of the Road Networks and a Country for West and Central African Countries." Paper
Typology of Response Measures." In Transporta- presented for the World Bank at the Senior Man-
tion Research Board. Transportation Research Re- agement Seminar on Planning, Financing, and
cord 1128. Washington, D.C. Managing Cost-Effective Road Maintenance Pro-
Faiz, Asif, and others. 1987. "The State of Roads in grams, Abidjan, June. Processed.
Developing Countries: Financial Needs and Cost- Mason, Melody, and Sergio Miquel. 1986. "Results
Effective Maintenance Options." In Japanese Min- of Road Maintenance Survey for Latin American
istry of Construction. "Text of the OECD Expert and Caribbean Countries." Washington, D.C.:
Meeting and Symposium on the Appraisal of the Transportation Department, World Bank. Pro-
Social and Economic Effects of Road Network cessed.
Improvements, Yokohama, May 26-28, 1987." Oakland, William H. 1985. A Survey of Arguments
Tokyo. Processed. and Practice in Earmarking Taxes. New Orleans,
*Harral, Clell. 1987. "Road Deterioration in Devel- La: Tulane University.
oping Countries: Organization and Management Paterson, William D. 0. 1987. Road Deterioration
of Road Maintenance." In Transportation Re- and Maintenance Effects: Models for Planning
search Board. Transportation Research Record and Management. Baltimore, Md.: Johns Hop-
1128. Washington, D.C. kins University Press.
*Harral, Clell, and James W. Eaton. 1986. "Improv- Robinson, Richard. 1985. Road Maintenance: Plan-
ing Equipment Management in Highway Author- ning and Management for Developing Countries.
ities in Developing Countries." Transportation Crowthorne, Berkshire: Overseas Unit, Transport
Department Discussion Paper, Transportation Is- and Road Research Laboratory.
sues Series, Report no. TRP2, World Bank, Wash- . 1987. "A View of Road Maintenance Eco-
ington, D.C. Processed. nomics and Policy in the Third World." Paper
*Harral, Clell, Ernesto E. Henriod, and Peter Gra- presented at the International Road Federation
ziano. 1986. "An Appraisal of Highway Mainte- Executive Conference on the Management of
nance by Contract in Developing Countries." Road Maintenance, Grand Canyon, February
Transportation Department Discussion Paper, 11-17. Processed.
Transportation Issues Series, Report no. TRP1, Robinson, Richard, Sydney Thriscutt, and Clell Har-
World Bank, Washington, D.C. Processed. ral. 1985. "The Road Maintenance Crisis in Afri-
Harral, Clell, and others. 1985. "Evaluating the Eco- ca: An Agenda for Action." Paper presented for the
nomic Priority of Highway Maintenance." Paper World Bank at the Senior Management Seminar on
presented for the World Bank at the Senior Man- Planning, Financing, and Managing Cost-Effective
agement Seminar on Planning, Financing, and Road Maintenance Programs, Abidjan, June. Pro-
Managing Cost-Effective Road Maintenance Pro- cessed.
grams, Abidjan, June. Processed. Rolt, John. 1981. Optimum Axle Loads of Commer-
Hicks, Norman, and Ann Kubisch. 1983. "Effects cial Vehicles in Developing Countries. Transport
of Expenditure Reductions in Developing Coun- and Road Research Laboratory Report 1002.
tries." CPD Working Paper, Country Programs Crowthorne, Berkshire: Transport and Road Re-
Department, World Bank, Washington, D.C. Pro- search Laboratory.
cessed. Sayers, M. W., T. D. Gillespie, and W. D. 0. Pater-
Hirschman, Albert. 1970. Exit, Voice, and Loyalty. son. 1986. Guidelines for Conducting and Cali-
Cambridge, Mass.: Harvard University Press. brating Road Roughness Measurements. World
Hoban, Christopher J. 1987. Evaluating Traffic Bank Technical Paper no. 46, Washington, D.C.
Capacity and Improvements to Road Geometry. Sayers, M. W., T. D. Gillespie, and C. A. V. Queiroz.
World Bank Technical Paper no. 74, Washington, 1986. The International Road Roughness Experi-
D.C. ment: Establishing Correlation and a Calibration

60
 Standard for M!easurements. World Bank Techni- U.S. Congressional Budget Office. 1983. Public
cal Paper no. 45, Washington, D.C. Works Infrastructure: Policy Consideration for the
*Smith, Graham, and Clell Harral. 1987. "Road 1980s. Washington, D.C.: U.S. Congress.
Deterioration in Developing Countries: Finan- U.S. Department of Transportation. 1976. America's
cial Requirements." In Transportation Research Highways 1776-1976. Washington, D.C.: Federal
Board. Transportation Research Record 1128. Highway Administration.
Washington, D.C. Watanatada, Thawat, Ashok M. Dhareshwar, and
Transport and Road Research Laboratory. 1987. Paulo Roberto S. Rezende Lima. 1987. Vehicle
"Maintenance Management for District Engi- Speeds and Operating Costs: Models for Road
neers." In Overseas Unit, Transport and Road Planning and Management. Baltimore, Md.: Johns
Research Laboratory. Overseas Road Note 1. Sec- Hopkins University Press.
ond ed. Crowthorne, Berkshire. Watanatada, Thawat, and others. 1987. The High-
Transportation Research Board. 1979. Consequences way Design and Maintenance Standards Model.
of Deferred Maintenance. Synthesis of Highway Vol. 1, Description of the HDM-III Model; vol. 2,
Practice no. 58. Washington, D.C. User's Manual for the HDM-III Model. Baltimore,
. 1986. "Technology Transfer to Transporta- Md.: Johns Hopkins University Press.
tion Agencies." Transportation Research Record Winfrey, Robley. 1968. Economic Analysis for High-
1080. Washington, D.C. ways. Scranton, Pa.: International Textbook Co.
U.K. Department of Transport. 1980. Policy for World Bank. 1981. The Road Maintenance Problem
Roads: England 1980. London: Her Majesty's Sta- and International Assistance. Washington, D.C.
tionery Office. . 1986a. The World Bank Atlas 1986. Wash-
U.K. National Development Office. 1985. Investment ington, D.C.
in the Public Sector Built Infrastructure-Report . 1986b. WVorldDevelopment Report 1986.
A: Roads and Bridges. London. Washington, D.C.
United Nations. 1983. Annual Bulletin of Transport
Statistics for Europe. New York.

61
The World Bank
Headquarters EuropeanOffice TokyoOffice
1818HStreet, N.W 66, avenued'1ena KokusaiBuilding
Washington,D.C.20433,U.S.A. 75116Paris,France 1-1Marunouchi3-chome
Telephone:(202)477-1234 Telephone:(1) 47.23.54.21 Chiyoda-ku,Tokyo100,Japan
Telex: WUI64145WORLDBANK Telex:842-620628 Telephone: (03) 214-5001
RCA248423WORLDBK Telex: 781-26838
Cable Address: INTBAFRAD
WASHINGTONDC

Relatedtitlesfromthe WorldBank
EconomicAppraisalof Thansport fProjects:A Manualwith CaseStudies
HansA.Adler.TheJohnsHopkinsUniversity Press.
Evaluating7hafficCapacityandlrmprovements toRoadGeometry.
ChristopherJ. Hoban.WorldBankTechnicalPaper7-4.
Guidelinesfor ConductingandCalibratingRoadRoughnessMeasurements
M.W.Sayers T.D.Gillespie,andW.D.0. Paterson.WorldBankTechnical
Paper46.
The HighwayDesignand MaintenanceStandardsSeries.TheJohns
HopkinsUniversity Press.
TheHighwayDesignandMaintenanceStandardsModeLThawat
Watanatadaandothers Vol.1:Descriptionof theHDM-I ModeL
Vol.2: UsersManualfortheHDMf-fIModel.
RoadDeteriorationandMaintenanceEffects:ModelsforPlanning
andManagement.WilliamD.0. Patersdn.
VehicleOperatingCosts:EvidencefromDevelopingCountries
AndrewChesherand RobertHarnison.
VehicleSpeedsandOperatingCosts:ModelsforRoadPlanningand
Management. ThawatWatanatada,AshokM.Dhareshwar,and Paulo
RobertoS. RezendeLima.
TheInternafionalRoadRoughnessExperiment:EstablishingCorrelation
anda CalibrationStandardforMeasuremenfsMichaelW.Sayers,Thomas
D.Gillespie,andCesarA.C.Queiroz.WorldBankTechnicalPaper45.
Labor-BasedConstructionPrograms: A PracticalGuideforPlanningand
ManagementBasilCoukisand others.OxfordUniversityPress,NewDelhi.
AlsoavailableinFrench.
TheRoadMaintenanceProblemandinternationalAssistance. Also
availablein Frenchand Spanish.
UrbanTransportAWorldBankPolicyStudy.AlsoavailableinFrenchand
Spanish.

ISSN 0258-2120 ISBN0-8213-1039-9