Public Disclosure Authorized
Public Disclosure Authorized
Public Disclosure Authorized
Public Disclosure Authorized
Policy Research
o9)
WePS
WORKING PAPERS
Transport
WesternAfricaDepartment
and Infrastructureand Urban Development
Department
The World Bank
June 1992
WPS921
RoadInfrastructure
and Economic
Development
Some DiagnosticIndicators
CesarQueiroz
and
SurhidGautam
The averagestockofpavedroadsper millioninhabitantsin highincome economiesis 59 times that in low-incomeeconomies.
And those roads are in better condition than the ones in lowincome economies.
of ideasamongBankstaffand
PolicyReerchWoikingPapis dissaninatetherrdingsof wotkinprogressandcacouragethicixchange
all othersimeestedin developrnatiswes.Thesepapas cany thenamesof theauthors,rcflectonlytheirviews.and shouldbe used and
citedacordingly.Thefindings,intepretations,andconclusionsar theauthors'own.Theyshouldnot be attibutedto theWorldBank,
its Board of Directors.its managanit, or any of its mnrnbercountries.
PolicyResearch
WPS 921
This paper- a joint productof the InfrastructureOperationsDivision,WesternAfricaDepartment,and
the TransportDivision,Infrastructureand Urban DevelopmcntDepartment- is part of a larger effortto
definethemacroeconomiclinkagesand impactofinfrastructure.Copiesofthepaperare availablefreefrom
the WorldBank, 1818H StreetNW, Washington,DC20433.PleasecontactMarieLaygo,room H3-151,
extension31261(June 1992,35 pages).
Queirozand Gautaminvestigatethe association
betweenper capitaincome and the magnitude
and qualityof road infrastructure.Theyadopt an
empiricalapproach,directlycomparingor
correlatinga country's income with selected
variablesassociatedwith existingroad networks.
densityof pavedroads (km/millioninhabitants)
variesfrom 170 in low-incomeeconomiesto
1,660in middle-incomeand 10,110in highincomeeconomies.That is, the averagedensity
of pavedroads in high-incomeeconomiesis 59
times that in the low-incomegroup.
Cross-sectionanalysisof data fromi98
countries,and time-seriesanalysisof U.S.data
since 1950,show consistentand significant
associationsbetweeneconomicdevelopment
(per capitaGNP) and road infrastructure(per
capitalength of pavedroad network).
Road conditionsalso seemto be associated
with economicdevelopment:The average
densityof paved roadsin good conditionvaries
from 40 km/millioninhabitantsin low-income
economiesto 470 in middle-incomeand 8,550in
high-incomeeconomies.
The data show that the per capitastock of
road infrastructurein high-incomeeconomiesis
dramaticallygreaterthan in middle-and lowincomeeconomies.For instance,the average
The empiricalinformationpresentedcan be
used as indicatorsof areasof weaknessor
strengthin a country's stockof road infrastructure.
The PolicyResearchWorkingPaperSeriesdisseminatesthe findingsof workunderwayintheBank.Anobjectiveoftheseries
;s to get these findingsout quickly, even if presentationsare less than fully polished.The findings,interpretations,and
conclusionsin these papersdo not necessarilyrepresentofficialBankpolicy.
Producedby the PolicyResearchDisseminationCenter
ROAD INFRASTRUCTUREAND ECONOMICDEVELOPMENT:
SOME DIAGNOSTICINDICATORS
By
Cesar Queiroz
Senior HighwayEngineer
Western Africa Department
World Bank
1818 H Street, NW
Washington,DC 20433 USA
Tel: (202) 473-5525
Fax: (202) 473-8249
and
Surhid Gautam
Research Assistant
Infrastructureand Urban DevelopmentDepartment
World Bank
1818 H Street, NW
Washington, DC 20433 USA
Tel: (202) 473-3743
Washington, D.C.
Road InfrastructureandEconomicDevelopment
SomeDiagnosticIdicators
by
Cesar ueiroz
and
SurhidGautam
Table of Contents
Introduction
2
How Roads InfluenceDevelopment
3
Sourcesof Data
3
Cross-SectionAnalysis
4
Time-SeriesAnalysisof U.S. Data
6
Comparisonof Cross-Sectionand llme-SeriesAnalyses
7
Comparisonof Road Supplyin the WorldEconomies
7
A Discussionof Causality
9
Conclusions
10
References
12
Figures
16
Table
25
Appendix
27
ROAD INERASTRUCTUREAND ECONOMIC DEVELOPMENT:
SOME DIAGNOSTICINDICATORS
by Cesar Queiroz and Surhid Gautam
An investigation of the association betweenper capita income and the magnitude and
quality of road infrastructureis carried out. The approach adopted is empirical in
that selected variables on existing road networksare directly compared or correlated
with a country's income. Cross-sectionanalysis of data from 98 countries, and timeseries analysis of U.S. data since 1950 showed consistent and significant associations
between economic development, in terms of per capita gross nationalproduct (GNP),
and road infrastructure, in terms of per capita length of paved road network. The
data show that the per capita stock of road infrastructurein high-income economiesis
dramatically greater than in middle and low-incomeeconomies. For instance, the
average density of paved roads (km/millioninhabitants)variesfrom 170 in lowincome economies to 1,660 in middle and 10,110 in high-incomeeconomies, the latter
being 5,800 percent higher than the low-incomegroup. Road condition also seems to
be associated with economic development: the average density of paved roads in
good condition (knmmillioninhabitants)variesfrom 40 in low-incomeeconomiesto
470 in middle and 8,5.50 in high-incomeeconomies. The empirical information
presented can be used as indicatorsof areas of weaknessesor strengths in a country's
road infrastructurestock.
Introduction
Road transport is an important sector of economic activity, especiallyin
developing countries, where it plays an essential role in marketingagricultural
products and providing access to health, education and agricultural inputs and
extension services. The impact of road transportationin developed regions is also
significant. As an example, in the United States it accounts for 15 per cent of the
Gross National Product (GNP) and 84 percent of all spending on transportation (1).
An efficient road system gives a country a competitiveedge in moving goods
economically. Conversely, lack of accessibilityor poor road conditions are barriers
to agriculture, industry and trade, and may hinder the entire developmenteffort.
Nevertheless, the contributionsof transport to national development may be difficult
to quantify in economic terms.
This paper presents information which can be used as indicators of
areas of weaknessesor strengths in a country's road infrastructure stock. The
approach adopted is empirical in that selected variables on existing road networks are
directly compared or correlated with a country's income. As pointed out by Owen
(2), comparisonsof income and road infrastructureare not meant to imply that a road
by itself is capable of developinga country or region, but that it is a necessary
element in the developmentprocess.
How Roads Influence Development
Transportationplays a multifacetedrole in the pursuit of development
objectives. Restriction of accessibilitylimits efficient factor mobility, and defers the
transfer of human and material resources to places where they can be employed most
productively. Conversely, transportationdevelopmenthelps to attain an efficient
distribution of population, industry and income.
Rural areas with low standardsof living are characteristicallythose
with inadequate methods of movingpeople and goods, probably because of deficient
access between villages.and markets, schools, medical, economic, administrativeand
social services which affect the day to day lives of rural people (3). Transportationis
an essential ingredientof almost everything man does to supply himself with the
necessities of life. Road transport is particularly important for developing countries,
where it provides about 80 to 90 percent of the total inland and/or border crossing
transport of people and goods. An effective road network can hasten progress in
agricultural and rural development, industryand trade, the viability of urban areas,
and the expansionof jobs, education and personal opportunity (4). The World Bank's
Long-Term Perspective Study (5) emphasizesthat although better market incentives
(especially related to prices and inputs) to farmers remain important factors in
agriculture, the effects of these would be blunted if the physical barriers and
economic costs of transportinggoods to and from markets remain high.
Sources of Data
Data used in this analysis were gathered from different sources. GNP
and population data come from the statisticalannexes of "Sub-SaharanAfrica - From
Crisis to SustainableGrowth" (5) and the World DevelopmentReport 1990 (6). The
data on road length, classificationand condition were compiled from different World
Bank reports (7, 8), "World Road Statistics 1985-1989"(9), Highway Statistics (10),
Statistical Abstracts of the United States (11), Annual Bulletin of Transport Statistics
for Europe (12), and World Transport Data (13).
The main variables includedin this study are defined as follows. Gross
National Product (GNP) is the measurementof a nation's total market value of the
final goods and services that are produced in the economy during a given time period,
normally one year. GNP per capita is a country's gross national product divided by
its population, henceforth denoted by PGNP. Spatial road density is a country's road
length per land area, and road density is per capita length of the road network.
Road conditionsare defined as in the World Bank policy paper on road
deterioration (5): (a) Good: paved roads substantiallyfree of defects and requiring
only routine maintenance,or unpaved roads needing only routine grading and spot
repairs; (b) Fair: paved roads having significantdefects and requiring resurfacing or
strengthening, or unpaved roads needing reshaping or resurfacing (regraveling)and
spot repair of drainage; and (c) Poor: paved roads with extensive defects and
3
requiring immediate rehabilitationor reconstruction,or unpaved roads neading
reconstruction and major drainage works.
The sample countries for which the required data was available are
listed in ^he Appendix. Naturally, the informationon the total length of paved and
unpaved road networksis deemed more reliable than the percentages in good, fair or
poor condition. Information on road condition from different countries is often
collected by different methods and on the basis of different definitions.
Cross-SectionAnalysis
We employ an empirical approach to explore the association between
road infrastructure and economic development. Different regression analyses were
carried out using GNP/capitaas dependentvariable and selected indicators of
magnitude and condition of road networks as independentvariables. Independent
variables used in the analyses included: a) spatial road density (i.e., road length per
land area) of paved and unpaved roads classified in good, fair or poor condition; and
(b) road density or per capita length (km/millionpopulation) of paved and unpaved
roads in good, fair or poor condition.
A sample of the relationshipsresulting from analyses of the data
described above, with per capita GNP as the dependent variable, is given in Table 1.
The most significantrelationship was between per capita GNP and density of paved
road network. Figure 1 shows this relationshipfor 98 developed and developing
countries in 1988, along with the scatter diagram. The resulting correlation equation
is:
PGNP = 1.39 x LPR
where PGNP is per capita GNP ($/inhabitant)and LPR is the per capita length of
paved roads (km/million inhabitants). The R squared value is 0.76, the number of
degrees of freedom is 97, and the t-statisticof the coefficient is 20.7. By changing
units in the above equation, one can show that there exists on the average $1.39 of
per capita GNP for each millimeterof paved road in a country. A less significant
regression equation (with an R squared value of 0.50) was obtained between per
capita GNP and the spatial density of paved roads (Figure 2): LGNP = 2.25 + 0.49
x LD, where LGNP is the logarithm of per capita GNP ($Iinhabitant),and LD is the
logarithm of spatial density of paved roads (km/l,000 square kn of land area); the
number of degrees of freedom is 96 and the t-statisticof the coefficientis 9.6.
The methodological,conceptual, and statisticalproblems
over cross-country studies of growth are well described by Levine and Renelt (14).
Notwithstandingthese problems, the relationshipdescribed above seems quite reliable
because of the relatively large sample size (98 countries) and its statistical
significance. It is also relatively consistent with the results of a time-series analysis
of U.S. data, as shown later in the paper.
4
The coefficient in the above equation (i.e. 1.39) can be used as a rough
indicator of the adequacy of paved roads stock in a country. Countries where the
ratio between per capita GNP and paved roads density is well above 1.39 are
relatively underendowedin terms of their stock of paved roads. Such is the case for
example of South Korea (ratio=15.5) and Bolivia (ratio=5.0). Conversely, countries
with low ratios can be viewed as relatively overendowedin terms of paved roads
stock. This is ttie case, for example, of Venezuela: with a ratio of 0.3, this country
should probably concentrate more in maintainingthe existing network instead of
paving new roads.
Road costs vary widely across countries and over time and depend on a
number of circumstances. Costs of new road constructioncan vary from less than
$50,000 a kilometer for a gravel road to more than $1 million a kilometer for a fourlane access-controlleddivided highway (7). If we assume that paved roads cost
$300,000/km (which is reasonable for developingcountries), the equation
PGNP= 1.39LPR can be used to show that on the average there exists $4.60 of per
capita GNP for each dollar invested on paved roads:
PGNP = 4.6 x I
where I is the investmentin paved roads (in dollars), i.e., the product of the lengtk of
paved roads and $300,000/km.
A rationale for the equation above is that "infrastructureinvestments
contribute to economic growth by increasing the productivity of other economic
inputs" (15). Although correlation does not imply causality, it is significantthat
economic developmentand road infrastructureare closely associated.
A fundamentalproblem with cross-countryestimationof supply
functions is the establishmentof the direction of causality. Supply functions
implicitly assume that such direction runs from road stock to output or productivity.
Therefore, causality is an interesting issue to be highlighted for future research: Does
an increase in road stock cause growth or is it the other way around? On a technical
point an argument could be made that increasing GNP results in less restricted
maintenanceand developmentbudgets and hence improved road infrastructure. To
prove that better roads lead to better GNP growth is not within the scope of this
paper. If further research proves this beyond any doubt, the above equation could be
interpreted as follows: an investment of $1 to expand the paved road network of a
country corresponds, on the average, to an increase of $4.6 (or about five times) of
the country's GNP. Lest the unwary reader interprets this as a rate of return of 460
percent, one should bear in mind that roads stock is just one of a large number of
inputs required to produce a certain level of output. Without causality, the
relationship PGNP = 4.6xI merely implies that for every $1 of a country's paved
road capital stock, we observe $4.6 of GNP (a brief discussionof causality is
presented later).
5
Moreover, the importance of unpavedroads should ,rot be
underestimated. For example, in many agriculturalareas, unpaved roads are the
feeder system critical to the marketingof the agricultural surplus which indirectly
supports the higher productive urb. l economy. The density of unpaved roads was not
included in the above equationsbecause of its high correlation with the density of
paved roads. It seems therefore prudent to interpret LPR as a proxy for a countty's
road stock, both paved and unpaved.
Time-Series Analysis of U.S. Data
A vast amount of historic data is available on the road network and
economyof the United States (10, 11). By carrying out a time-series analysis of U.S.
data from 1950 to 1988, we found a significantpositive relationshipbetween per
capita GNP (PGNP, in $1,000/inhabitant,using 1982 constant dollars) and density of
paved roads (LPR, in km/i,000 inhabitants):
PGNP = -3.39 + 1.24LPR
with an R squared value of 0.93; the number of degrees of freedom is 37, and the tstatistic of the coefficient is 21.4 (Figure 3). The intercept (i.e., -3.4) in the above
equation is difficult to interpret. However, a null GNP is well beyond the inference
space. Moreover, if we force the equation through the origin, the resulting regression
equation is still signiiicant: PGNP = 0.97LPR, with an R squared of 0.88.
An interesting exercise consists in running regressions between PGNP
and LPR using different time lags: we found the highest correlation existed when
PGNP for a given year was associated with LPR four years earlier (Figure 4). This
seems to indicate that paved roads had an effect on GNP, but there was a time lag of
about four years between constructionand ultimate impact. This four-year time lag is
in broad agreement with the "half a decade" lag period observed by Aschauer (16).
Aschauer has shown that productivity (i.e., output per unit of private capital and
labor) is positively related to governmentspending on infrastructure, including roads.
Analyzing data from the United States for the period 1949 to 1985, he observed that
underinvestmentin infrastructure started in about 1968, and the effects of
deterioration became evident half a decade later, when a productivity slump began in
the U.S.
It should be noted, however, that the above result was obtained for only
one country, and using only one independentvariable in the equation. This is an area
specificallyrecommendedfor further research, in that similar exercises could be
carried out for other countries with inclusion of additionalexplanatory variables in the
equations.
6
Comparisonof Cross-Sectionand Time-SeriesAnalyses
It is interestingto comparethe equationsresultingfrom the crosssectionanalysisof data from 98 countries(circa 1988)and from the time-series
analysisof the U.S. data (1950to 1988). The time-seriesequationPGNP = -3.4 +
1.24LPRwas derivedwithconstant1982dollars. To makeit comparablewith the
cross-sectionalequation,it shouldbe expressedin 1988constantdollarstakinginto
accountthe changein the GNP implicitprice deflatorbetween1982and 1988.i.e., a
factorof 1.213(11). The resultingequationis:
PGNP88= -4.1 + 1.50 x LPR
wherePGNP
and LPR is the per
8 8 is real per capitaGNP (1988$1,000/inhabitant)
capitalength(or density)of paved roads (km/thousandpopulation).
The inferencespacesfor both equationscan be approximatelydefined
by: (a) cross-sectionalanalysis: paved road densitybetween60 and 20,000
km/millionpopulation;and (b) time-seriesanalysis: pavedroad densitybetween
8,000and 20,000 km/millionpopulation. Figure5 depictsthe two equations
accordingto their inferencespace. As can be seenin the figure, thereis relatively
goodconsistencebetweenbothequations.
Comparsonof Road Supplyin the WorldEconomies
A comparisonbetweenthe supplyand conditionof pavedroad networks
in 98 developingand developedcountriesis shownin Figure 6. The countrygroups
in the figure are definedas (6):
(a) Low-incomeeconomiesare those witha GNPper capitaof $545or
less in 1988;
(b) Middle-income
economiesare thosewith a GNPper capitaof more
than $545but less than $6,000in 1988;and
(c) High-incomeeconomiesare those witha GNPper capitaof $6,000
or morein 1988.
For the analysesdescribedin this paper,data wasavailablefor 42 lowincomeeconomies(averageper capitaGNP of $320); 43 middle-income
economies
(averagePGNP=$1720/capita);and 13 high-incomeeconomies(average
PGNP=$ 17,420/capita).
As shownin Figure6, the supplyof road infrastructurein high-income
economiesis dramaticallyhigherthanin middleand low-incomeeconomies. For
instance,the averagedensityof pavedroads (km/millioninhabitants)varies from 170
in low-incomeeconomiesto 1,660(plus 876percent)in middleand 10,110in high7
incomeeconomies,the latter being5,%q0percenthigherthan the low-incomegroup.
Road conditionis also associatedwith economicdevelopment:the averagedensityof
pavedroads in goodcondition(km/millioninhabitants)varies from 40 in low-income
economiesto 470 in middleand 8,550 in high-incomeeconomies(an increaseof
21,000percentover the low-incomegroup).
In the particularcase of Africa,there is a similartrendbetweenlow
and middle-income
economies,as shownin Figure 7. Whilethe increasein average
per capitaGNP betweenthe two countrygroupsis 220 percent,the per capita length
of pavedroads in goodconditionincreasesby about370 percentwith the increasein
income.
The results aboveseemto indicatethat economicdevelopmenthas a
link withpaved roadsdensity,and also to the maintenancestandardsof theseroads.
A similartrendexistsfor unpavedroads,sincethere is highcorrelationbetweenthe
extentof a country'spavedand unpavedroad networks.
The limitedresourcesdevotedto the upkeepof road networksin
developingcountriesin the last decade,togetherwith the growthof heavyfreight
traffic,have createda large backlogof road maintenanceand rehabilitationneeds. In
severalcountriesmanykilometersof roads havede.erioratedfrom good to fair and
from fair to poor condition. It is not exceptionalfor sectionsof maintrunk roads to
have lost mostor all of their blacktop, thus causinga shrinkageof a country's
passableroad network. Althoughmanyother factorsare involved,severalcountries
whereGNP per capitahas decreasedin recent yearshave also facedsignificant
deteriorationin their road networks. This trendis illustratedin Figure 8, which
showsa declinein real per capitaGNP and road conditionbetween1984and 1989for
severalAfricancountries.
Conversely,severalcountriesthat were able to improvetheirroad
infrastructurein the sameperiodhad also an increasein real incomeper capita
(Figure9). Ghanais a good example: From 1984to 1989,its per capitaGNP
increasedby 11 percent--from$350annuallyto $390; in the sameperiod, the density
of pavedroads in goodconditionexpandedby 102percent,from 56 to 113km per
millioninhabitants(Figure9). On the basisof the findingsdescribedabove,
improvedroad infrastructurein Ghanais likelyto contributeto furthereconomic
growth.
Regardingtt., nacroeconomic
linkagesof infrastructure,authorssuch
as Ingram (15)assert that the conceptuallink betweeninfrastructure(includingroads)
and the supplyside of the economyis as follows: (a) reductionsin the stock of
infrastructurecapitalcan shiftthe productionpossibilityfrontierinwardand reduce
the economy'spossibleoutput;and (b) increasesin infrastructurecapitalcan shift the
productionpossibilityfrontieroutwardand providea sourceof growthfor the
economy. This linkageis in line with the data shownin Figures8 (reductionin both
8
per capita GNP and density of paved roads in good condition)and 9 (increase in both
per capita GNP and density of paved roads in good condition).
A Discussionof Causality
Assessingthe impact of road infrastructure on economic performance is
not straightforward because many other factors are involved. As we mentioned
earlier, direction of causation between changes in income and changes in road
infrastructure are not clear cut. One could argue that causation in the equations
described in this paper could run in either direction. However, there are .*jme
indications that roads should precede development,of which a few examples are:
(a) In estimating the aggregate supply response of agriculture,
Chhibber (17) showedthat both price and non-pricevariables have significanteffect.
Although he did not explicitly consider a proxy for road stock in his analyses, it was
implied that non-price variables included transport and communicationfacilities.
(b) Binswangex(18) found that the lack of roads is a significant
constraint on the supply response of agriculture.
(c) In India, a Central Road ResearchInstitute study by Dhir, Lal and
Mfital(19) has shown that literacy, agriculturalyield and health care increase with
road density.
(d) Shah (20) used a restricted equilibriumframework to estimate the
contribution of public investmentin infrastructure to private sector profitability in
Mexico. He concluded that a policy emphasis should be to upgrade the public
infrastructure (includingroads) so that scale economies could be exploited in the
future.
(e) The linkage between roads and developmenthas also been
acknowledgedby country leaders. As an example, President Bush has asserted that
the interstate highway system fueled developmentin the U.S. for a generation, uniting
the states as never before - economically,politicaily, socially (21).
(f) Hirschman (22) pointed out that highwayconstruction can be
conceivedas the laying down of a "prerequisite" for further development. As such, it
permits and invites, rather than compel, other activitiesto follow suit. This is in line
with Owen's (2) assertion that comparisonsof income and road infrastructure are not
meant to imply that a road by itself is capable of developinga country or region, but
that it is a necessary element in the developmentprocess.
(g) Using US data in the period 1949 to 1985, Aschauer (16), has
shown that productivity (i.e., output per unit of private capital and labor) is positively
related to government spending on infrastructure, including roads. He also observed
that underinvestmentin the U.S. infrastructurestarted in about 1968, and the effects
9
of deteriorationbecame evident half a decade later. In a different paper, Aschauer
(23) offered several checks on the direction of causation, concludingthat increases in
public capital stock lead to higher total factor productivity, which is a proxy for per
capita income.
(h) Using 1965 data from 47 less developedcountries and 19
developed countries, Antle (24) demonstratedthe importanceof transportationand
communicationfacilities for raising aggregate agriculturalproductivity.
(i) An analysis of the economic rates of return of World Bank financed
projects in the period 1968 to 1984, carried out by Israel (25), indicates that
transportationinvestments, particularly roads, are among the most productive.
Therefore, the notion that road infrastructure is a necessary element in
the developmentprocess is supportedby several pieces of research. However, many
factors can influence the impact of roads on income. In particular, an exploration of
the linkages between policy distortionsand the actual outcome of infrastructure
investments, carried out by kauffman (26), concludedthat a distorted policy
environmentreduced significantlythe ex-post return of the investments. A good
example of policies that would probably increase the impact of road investmentson
productivity was given by Small, Winston and Evans (27). Their policy
recommendationsinclude a set of pavement-weartaxes for heavy trucks, a set of
congestion taxes for all vehicles, and a program of optimal investmentsin road
durability. Such policies are based on two economic principles: efficient pricing to
regulate demand for highway services and efficient investment to minimizethe total
public and private cost of providing them (27).
Conclusions
The analyses in this paper show that there is a statistically significant
relationship between road infrastructure and economic developmenton a worldwide
basis: cross-sectionanalysis of data from 98 countries (circa 1988), and time-series
analysis of U.S. data between 1950 and 1988 showed significantrelationshipsbetween
per capita gross national product (PGNP) and density (i.e., per capita length, LPR) of
paved road network. Moreover, there is relatively good consistencybetween the
regression equationsfrom cross-sectionand time-seriesanalyses, when compared
according to their respective inference space. Becauseof the high correlation between
the densities of paved and unpaved roads, LPR should be interpreted as a proxy for a
country's road stock, both paved and unpaved.
The per capita stock of road infrastructurein high-incomeeconomies is
dramaticallygreater than in middle and low-incomeeconomies. For instance, the
average density of paved roads (kmn/million
inhabitants)varies from 170 in lowincome economies to 1,660 (plus 876 percent) in middle and 10,110 in high-income
economies, the latter being 5,800 percent higher than the low-incomegroup. Road
conditionalso seems to be associated with economic development: the average
10
density of paved roads in good condition (km/million inhabitants)varies from 40 in
low-incomeeconomies to 470 in middle and 8,550 in high-incomeeconomies. There
is also a clear contrast between road infrastructureand income in low and middleincome economies in Africa: while the differencein average per capita GNP between
the two country groups is 220 percent, the density of paved roads in good condition
varies by about 370 percent from one group to the other, using 1989 data.
Causality is an issue highlightedfor future research: Does an increase
in road stock cause growth or is it the other way around? Assessing the impact of the
supply and quality of road infrastructureon economic performance is a complex area
of research with potentiallyimportant implicationson the intemational infrastructure
lending strategy to developingcountries.
Acknowledgements
The authors benefittedfrom comments of James Wright, Jean Doyen,
Peter Morris, Charles L. Wright, Manuel Santos, Harold Young, Hiroshi Ueno, John
Riverson, Antoine G. Hobeika, Merron L. Latta, Jeremy Warford, Ajay Chhibber,
Arturo Israel, Peter Cook, John Roome and Mohsen Fardi.
11
Rekences
1.
Roads to Serve the Nation-TheStory of Road Developmentin the United
States. U.S. Federal Highway Administration,Publication No. FHWA-PL-89024, Washington,D.C., 1989.
2.
Wilfred Owen. Transportationand World Development. The Johns Hopkins
University Press, Baltimore, 1987.
3.
Riverson, J. D. N., and S. Carapetis. The Potential of IntermediateMeans of
Transport in ImprovingRural Travel and Transport in Sub-SaharanAfrica.
Sub-SaharanAfrica Transport Program, World Bank, Washington, D.C.,
1991.
4.
World Highways. InternationalRoad Federation, Vol. XLI, No. 8,
Washington, D.C., November/December1990.
5.
Sub-SaharanAfrica - From Crisis to SustainableGrownth.A Long-Term
PerspectiveStudy, World Bank, Washington, D.C., 1990.
6.
World DevelopmentReport 1990, World Bank, Washington,D.C., June 1990.
7.
Road Deteriorationin DevelopingCountries: Causes and Remedies, The
World Bank, Washington,D.C., 1988.
8.
UNTACDA U, Roads Sub-sectorWorking Group, "StrategyPaper", Africa
Technical Department, infrastructureDivision, The World Bank, Washington,
D.C., Dec. 1990.
9.
World Road Statistics 1985-1989,International Road Federation, Washington,
D.C., 1990.
10.
HighwayStatistics, U.S. Department of Transportation,Federal Highway
Administration,Washington,D.C. (Different Issues).
11.
StatisticalAbstracts of the United States 1991: The National Data Book.
Bureau of the Census, U.S. Department of Commerce, Washington,D.C.
(Different Issues).
12.
Annual Bulletin of Transport Statisticsfor Europe, United Nations, New York,
1990.
13.
World Transport Data, International Road Transport Union, Geneva, 1990.
12
14.
Levine, Ross, and David Renelt. 1991. Cross-Country Studies of Growth and
Policy: Methodological,Conceptual, and Statistical Problems. Working
Paper WPS 608, World Bank, Washington, D.C.
15.
Ingram, Gregory K. Note on the MacroeconomicLinkages of Infrastructure.
World Bank, Washington, D.C., October 1989.
16.
Aschauer, David A. InfrastructureExpendituresand Macro Trends. In
Proceedings of the Africa Infrastructure Symposium,World Bank,
Washington, D.C., 1989.
17.
Chhibber, Ajay. The Aggregate Supply Response: A Survey. In Simon
Commander, ed., Structural Adjustmentand Agriculture: Theory and Practice
in Africa and Latin America. Overseas DevelopmentInstitute, London, 1989.
18.
Binswanger, Hans. The Policy Responseof Agriculture. In Proceedings of
the World Bank Annual Conferenceon DevelopmentEconomics 1989. World
Bank, Washington, D.C., 1990.
19.
Dhir, M., N. Lal and K. Mital. The Developmentof Low-VolumeRoads in
India. Fourth InternationalConference on Low-VolumeRoads, Transportation
Research Board, TRR 1106, Vol. 2, Washington, D.C., 1987.
20.
Shah, Anwar. Dynamicsof Public Infrastructure, IndustrialProductivity and
Profitability. World Bank, Washington, D.C., 1990 (Forthcoming,The
Review of Econiomicsand Statistics, Harvard University).
21.
WashingtonPost, The. What Path Lies Ahead for U.S. Highways? No. 183,
114th Year, June 6, Washington,D.C., 1991.
22.
Hirschman, Albert 0. The Strategy for EconomicDevelopment. Yale
University Press, Inc., New Haven, 1958.
23.
Aschauer, David A. Is Public Expenditure Productive? Journal of Monetary
Economics 23, Elsevier Science PublishersB.V. (North-Holland),Amsterdam,
1989, p. 177-200.
24.
Antle, John M. Infrastructureand Aggregate Agricultural Productivity:
International Evidence. EconomicDevelopmentand Cultural Change, Vol.
31, No. 3. The University of Chicago Press, April 1983.
25.
Israel, Arturo. InfrastructureFramework Paper. Infrastructureand Urban
DevelopmentDepartment, World Bank, Washington,D.C., 1991.
13
26.
Kauffman, Daniel. Determinantsof the Productivityof Projects in Developing
Countries: Evidence from 1,200 Projects. BackgroundPaper to World
DevelopingReport 1991, World Bank, Washington,D.C., 1991.
27.
Small, Kenneth A., Winston, Clifford and Evans, Carol A. Road Work: A
New Highway Pricing and InvestmentPolicy. The BrookingsInstitution,
Washington, D.C., 1989.
14
List of Pigures
1.
Relationshipbetweenper capitaGNP and paved roadinfrastructure
2.
Relationshipbetweenper capitaGNP and spatialdensityof paved roads
3.
Relationshipbetweenper capitaGNP and paved roaddensityin the U.S.
4.
Correlationbetweenper capitaGNPand pavedroads in the U.S. using
differenttime lags
5.
Comparisonbetweencross-sectionand time-seriesanalyses
6.
Comparisonbetweenthe extentand conditionof pavedroad networksin 98
developingand developedcountries
7.
Comparisonbetweenthe extentand conditionof pavedroad networksin low
economiesin Africa
and middle-income
8.
Sampleof Africancountrieswith deterioratingper capitaGNP and road
conditionbetween1984and 1989
9.
Sampleof Africancountrieswith improvedper capita GNPand road condition
between1984and 1989
List of Tablea
1.
Sampleof relationshipsbetweenper capitaGNP (dependentvariable)and road
infrastructurevariables
Appndix
List of countrieswith data availableon population,per capitagross nationalproduct,
lengthof pavedand unpavedroadnetworks,and road condition. The vast majorityof
the data was obtainedfor 1988. However,for a few countries1988data was not
available;then 1987or 1989data was used. Countriesare sortedby per capitaGNP.
The appendixincludesthree annexes: (1) data on population,per capita GNP and
paved and unpavedroad density;(2) data on landarea, per capitaGNP and pavedand
unpavedroad spatialdensity;and (3) U.S. data usedin the time-seriesanalysis.
15
Relationshipbetween per capita GNP
and PavedRoadDensity (circa 1988)
GNP/Capita,$
50000
5000
98 countries
R squared: 0.76
PGNP 1.39x LPR
500
50
50
1
500
5000
50000
PavedRoads,km/millioninhabitants
FIGURE1
Relationshipbetweenper capita GNPand
PavedRoadSpatialDensity (circa 1988)
GNP/Capita,$
100000-
10000
'
a
a
1000 -
98 countries
R squared = 0.50
PGNP* 178 Density
100-
10
-
0.1
1
1
10
I
111111
I
100
Il
11
1l
1000
Pavedroads, km/1,000 sq km
FIGURE2
I
II11111
10000
RelationshipBetween Real per Capita GNP
and Paved Road Density in the U.S.
18 -
GNP/Capita in '82 Constant Dollars (000)
16 -
14I-'
PGNP--3.41.24xLPR
R squared:0.93
1950 to 1988
~12-
10
8-
6
48
I
9
l
10
l
11
12
13
14
Paved Road Density (km/thousand pop)
FIGURE 3
15
CorrelationBetweenReal per capita GNP
and LaggedPavedRoadDensity in the
U.S.for the EquationPGNP=aLPR
0.92
R squared
0.90
. .. .......
0.88 - ----
-------
0.84 1
-2
-1
.
.......
0
1
2
3
Number of Years Lagged
FIGURE4
4
5
ComparisonBetween Cross Section
and Time Series Analyses
Real GNP/Capita, 1988 $1,000
30 -
98 Countries
R squared: 0.76
25
20
o15- -
U.S.A.
1950-
//
10
0-
0
1988
R squared: 0.93
I
I
5
10
15
20
25
Paved Road Density, km/thousand pop
FIGURE5
30
AverageRoadDensity in Low, Middle,
and High IncomeEconomies
100000-
GNP/capita, $
Density, km/mil inhab
-100000
17420 10110 8660
10000
10000
1660
1720
1000
470
1000
170
320
4010
100
10
PGNP< $545
$545-$6,000 PGNP> $6,000
GNP/Capita
Paved Roads
Emi Good Paved Roads
FIGURE6
GNP and RoadDensity in Low and Middle
IncomeEconomiesin Africa
1200
GNP/Capita, $
Roads, km/mil inhabitants
1200
1000
938
951
1000
800
8 00
600 --
600
400
386400
200 -200
0
0
PGNP($545
GNP/Capita .
PGNP4$545
PavedRoads EZ Good PavedRoads
FIGURE7
ComparisonBetweenReal GNPand Good
PavedRoadper MillionInhabitants
(1984 and 1989)
$/capita
Congo 1984
1989
km/mil. inhab
1140
940
Zimbabwe 1984
422
287
760
1989
488
650
Zambia 1984
360
470
1989
344
390
Gambia, The 1984
280
260
172
1989
220
Tanzania 1984
210
1989
57
130
11111
I
I
10000
136
(fI~I
I
I
1000
GNP/Capita
FIGURE8
36
I
100
I
I
I
l111111
I
100
ME Good Paved Roads
I liltI
1000
ComparisonBetween Real GNP and Good
PavedRoads per Million Inhabitants
(1984 and 1989)
$/capita
Botswana 1984
1989
km/mil. inhab
960 -
1770
1600 -
1802
Mauritania 1984
450
1989
500
446
Ghana 1984
350
1989
390
Zaire 1984
1989
10000
289
56
113
140 E
14
260
1000
GNP/Capita
FIGURE 9
31
100
500
5000
Good Paved Roads
Table 1. Summary of Statistical Models Relating Income and Road Infrastructure
Equation
Estimation
No. of
method observ. Constant
1
OLS
98
732.8
2
OLS
98
0
3
OLS
98
2.6
4
OLS
98
0
5
OLS
83
564
6
OLS
83
0
7
OLS
39
-3.39
8
OLS
39
0
9
OLS
38
-2.9
10
OLS
38
0
1.1
OLS
37
-2.5
12
OLS
37
0
13
OLS
41
258
14
OLS
41
0
Xi
0.01
(0.03)
0.32
(1.33)
1.39
(20.74)
1.39
(21.06)
X2
1.67
(5.38)
1.39
(4.63)
1.18
(7.37)
1.63
(10.1-8)
1.24
(21.4)
0.97
(88.18)
1.22
(21.4)
0.99
(92.5)
1.2
(21.1)
1.0
(100)
0.44
(2.6)
0.74
X3
X4
X5
R
squared
Remarks
0.82
World
0.81
World
0.76
World
0.76
World
0.39
World
0.18
World
0.93
USA
0.88
USA
0.93
USAlag 1 yr.
0.89
USAlag 1 yr.
0.92
USAlag 2yrs.
0.90
USA lag2 yrs.
0.14
Africa
0.03
Africa
Table 1. Summary of Statistical Models Relating Income and Road Infrastructure
Equation
Estimatfon No.of
method observ. Constant
Xi
X2
X3
X4
R
X5 squared
Remarks
0.16
Africa
0.23
Africa
0.17
Africa
0.32
Africa
0.38
Africa
0.34
Africa
0.59
World
0.53
World
(6.2)
15
OLS
41
335
16
OLS
41
195
17
OLS
41
0
18
OLS
21
300
19
OLS
21
300
20
OLS
21
0
21
OLS
98
1.3
22
OLS
98
3.7
_ _ _ _ _ _ _ _ _ _ __ __ __ _ _ _ __ _ __ __ _ _ _ _ _ _ _
Notes:
Xi = Pavedroads(km/million
inhabitants)
X2= Goodpavedroads(km/millioninhabitants)
X3 = Unpavedroads(km/million
inhabitants)
X4= PavedRoads(kmf000sq.km)
X5= GoodPavedRoads(kmOOO
sq. km)
OLS= OrdinaryLeastSquare
t statisticsin parenthesis
0.59
(2.8)
0.21
(3.5)
0.29
(7.25)
0.60
(1.5)
66
(.99)
6.7
8.7
_ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _(1
3.0)
1
_
_
_
_
ANNEX1
DATA ON POPULATION,GNP/CAPITA,PAVEDAND UNPAVEDROADS
PAVEDROADKM ---(per millioninhabitants)
LENGTH
GOOD
(km)
(km)
-----
Countries
-4
Ethiopia
Chad
Guinea Bissau
Tanzania
Bangladesh
Malawi
Mozambique
Zaire
Equa.Guinea
Myanmar
Nepal
Madagascar
BurkinaFaso
Gambia,The
Mali
Burundi
Uganda
Nigeria
Somalia
Zambia
Niger
SierraLeone
Djibouti
Rwanda
China
India
Population
(million)
GNP/
Capita
47.4
5.4
0.9
24.7
102.1
8.0
14.9
33.4
0.4
40.0
18.0
10.9
8.5
0.8
8.0
5.1
16.2
110.1
5.9
7.6
7.3
3.9
0.4
6.7
1,088
816.0
130
150
160
160
170
170
170
170
180
180
180
190
210
220
230
240
280
290
290
290
300
300
320
320
330
340
84
56
604
146
61
276
343
84
1,118
210
139
477
177
638
308
198
111
310
467
724
379
196
1,030
145
159
150
40
0
236
36
9
155
41
32
303
0
56
267
42
140
194
115
11
208
243
289
227
122
525
59
16
30
---FAIR
(km)
35
6
157
44
24
105
257
19
560
105
49
129
87
294
96
50
70
16
154
217
87
18
393
85
111
68
UNPAVEDROADKM ---(per millioninhabibants)
LENGTH
GOOD
(km)
(k,m)
200
1,296
2,324
794
n.a.
930
574
1,740
1,608
n.a.
n.a.
486
851
1,000
1,230
605
269
228
788
1,994
538
932
1,800
710
na.
n.a.
94
n.a.
140
79
n.a.
74
34
765
483
n.a.
n.a.
131
n.a.
320
234
121
0
na.
32
598
129
75
918
135
n.a.
n.a.
FAIR
(km)
62
n.a.
140
238
n.a.
707
252
504
675
n.a.
n.a.
173
681
390
160
345
382
23
79
698
156
345
685
328
na.
n.a.
ANNEXI (continued)
DATAON POPULATION,
GNP/CAPITA,
PAVEDAND UNPAVEDROADS
-----
Countries
Pakistan
Comoros
Kenya
Togo
CA.R.
Haiti
Benin
Ghana
Lesotho
SriLanka
Guinea
Indonesia
Uberia
Mauritania
Sudan
Yemen
Senegal
Zimbabwe
Philippines
Morocco
Honduras
Swaziland
PapuaNewGuinea
Egypt
Coted'lvoire
Thailand
Populaffon
(million)
106.3
0.4
22.4
3.4
2.9
6.3
4.4
14.0
1.7
16.6
5.4
175.0
2.4
1.9
23.8
10.2
7.0
9.3
53.4
21.4
4.2
0.7
3.4
45.9
11.2
50.0
GNP/
Capita
350
370
370
370
380
380
390
400
420
420
430
440
450
480
480
522
650
650
660
670
700
700
710
720
770
860
PAVED
ROADKM---(permillioninhabitants)
LENGTH
GOOD
(km)
(km)
227
1,228
280
441
152
96
236
429
355
536
241
157
232
789
98
951
540
1,370
266
692
384
984
214
329
355
560
41
528
90
176
46
0
61
120
188
54
120
47
197
458
26
370
151
370
82
138
192
344
73
128
266
280
---FAIR
(km)
114
650
146
97
53
96
118
191
103
161
0
47
30
237
42
527
173
41
144
305
165
344
96
108
89
168
UNPAVED
ROADKM---(permillioninhabitants)
LENGTH
GOOD
(km)
(km)
n.a.
1,113
1,681
388
3,055
177
545
602
1,169
n.a.
1,056
n.a.
1,412
316
246
n.a.
929
2,467
n.a.
n.a.
n.a.
2,954
n.a.
n.a.
982
n.a.
n.a.
278
1,109
78
Z078
16
60
102
187
n.a.
n.a.
n.a.
212
51
49
n.a.
65
1,234
n.a.
na.
n.a.
1,773
na.
n.a.
334
n.a.
FAIR
(kn)
na.
425
252
39
489
48
190
241
666
n.a.
n.a.
na.
1,059
104
49
nma.
195
740
n.a.
n.a.
n.a.
1,093
n.a.
n a.
638
n.a.
ANNEX1 (continued)
DATA ON POPULATION,GNP/CAPITA,PAVEDAND UNPAVEDROADS
-----
Countries
Congo
DominicanRep.
Peru
Botswana
Cameroon
Bolivia
Belize
Ecuador
Jamaica
Guatemfala
CostaRica
Tunisia
Turkey
Colombia
SyrianArab. Rep.
Chile
Brazil
Mauritius
Poland
Portugal
Panama
Mexico
Malaysia
Yugoslavia
Argentina
Romania
Population
(million)
2.1
6.1
18.2
1.2
11.2
6.2
0.2
9.1
2.2
7.7
2.5
7.0
53.8
28.4
11.6
11.8
132.6
1.1
37.9
10.2
2.1
76.8
16.9
23.0
30.1
22.7
GNPI
Capita
910
970
1,000
1,010
1,010
1,099
1,150
1,150
1,150
1,160
1,190
1,270
1280
1,390
1680
1,700
1,720
1,800
1860
1,970
1,980
2,040
2110
2,120
2,230
2,290
PAVEDROADKM ---(per million inhabitants)
LENGTH
GOOD
(km)
(km)
593
407
394
1,967
290
218
2,210
371
1,984
395
1,218
1,306
846
339
1971
813
763
1,509
5804
1,755
1,473
843
1923
1,593
899
617
297
212
95
1,848
110
46
707
197
198
28
268
718
n.a.
142
n.a.
342
229
1,434
n.a.
877
530
716
n.a.
478
315
426
---- UNPAVED ROADKM---(per million inhabitants)
FAIR
LENGTH
GOOD
(km)
(km)
(km)
71
41
95
78
78
105
1,061
71
1,448
198
353
470
n.a.
125
n.a.
320
75
n.a.
526
796
84
n.a.
653
189
130
4,550
23
n.a.
4,870
2,627
1,477
8,250
530
1,818
77
1,522
n.a.
5113
563
524
1,947
n.a.
127
3711
n.a.
1,942
n.a.
451
n.a.
n.a.
n.a.
1,729
3
n.a.
2,192
420
295
1,320
355
n.a.
23
274
n.a.
n.a.
203
n.a.
175
n.a.
115
n.a.
n.a.
117
n.a.
n.a.
n.a.
n.a.
n.a.
FAIR
(km)
1,229
7
n.a.
925
1,497
532
6,270
116
1,073
14
928
n.a.
n.a.
276
n.a.
1,499
n.a.
6
n.a.
n.a.
1,262
n.a.
n.a.
n.a.
n.a.
n.a.
ANNEX1
DATA ON POPULATION,GNP/CAPITA,PAVEDAND UNPAVEDROADS
-----
Countres
Algeria
Uruguay
Gabon
Venezuela
Trinidadand Tobago
Cyprus
Korea,Rep.
SaudiArabia
Oman
Australia
UnitedKingdom
Italy
Belgium
Netherlands
Austria
France
Denmark
U.S.A.
Japan
Switzerland
Population
(million)
23.8
3.1
1.1
18.8
1.2
0.7
42.0
14.0
1.1
16.5
57.1
57.4
9.9
14.8
7.6
55.9
5.1
246.3
122.6
6.6
GNP/
Capita
2,360
2,470
2,970
3,250
3,350
3,590
3,660
6200
6,490
12,340
12,810
13,330
14,490
14,520
15,470
16,090
18,450
19,840
21,020
27,500
PAVEDROADKM ---(per millioninhabitants)
LENGTH
GOOD
(km)
(km)
1,365
2,079
636
10,063
1,733
4,197
236
2414
2,993
25,746
6,170
5,259
12,443
6,856
14,092
14,402
13,856
14,172
6,008
10,766
546
541
191
4,025
1,248
1,595
165
na.
1,975
21,883
5,244
4,470
10,577
5,828
11,978
12,242
11,778
12,047
5,107
9,151
Sources:World DevelopmentReport,RoadDeteriorationin DevelopingCountries,
IntemationalRoadFederation,and Staff AppraisalReports.
---FAIR
(km)
437
1,227
191
3,824
329
1,595
59
na.
599
3,862
925
789
1,866
1,028
2,114
2,160
2,078
2,126
901
1,615
UNPAVEDROADKM ---(per million inhabitants)
LENGTH
GOOD
(km)
(klm)
n.a.
915
4,182
1,301
n.a.
n.a.
n.a.
4111
n.a.
25,951
n.a.
na.
518
935
na.
na.
n.a.
11,135
2,999
n.a.
n.a.
421
1,338
299
n.a.
na.
n.a.
na.
na.
na.
n.a.
na.
na.
na.
na.
n.a.
na.
na.
na.
na.
FAIR
(km)
na.
430
1,255
781
na.
n.a.
na.
nA.
na.
na.
na.
n.a.
na.
na.
na.
na.
n.a.
na.
na.
na.
ANNEX2
DATA ON POPULATION,GNP/CAPITA,AREA, PAVEDAND UNPAVEDROADS
-----
Countries
I
Ethiopia
Chad
GuineaBissau
Tanzania
Bangladesh
Malawi
Mozambique
Zaire
Equa.Guinea
Myanmar
Nepal
Madagascar
BurkinaFaso
Gambia,The
Mali
Burundi
Uganda
Nigeria
Somalia
Zambia
Niger
SierraLeone
Djibouti
Rwanda
China
India
Population
(million)
47.4
5.4
0.9
24.7
102.1
8.0
14.9
33A4
0.4
40.0
18.0
10.9
8.5
0.8
8.0
5.1
16.2
110.1
5.9
7.6
7.3
3.9
0.4
6.7
1,088
816.0
GNP/
AREA
Capita (thou. km2)
130
150
160
160
170
170
170
170
180
180
180
190
210
220
230
240
280
290
290
290
300
300
320
320
330
340
1,222
1,285
36
945
144
119
802
2,345
28
677
141
587
274
11
1,240
28
236
924
638
753
1,267
72
22
26
9,561
3,228
PAVEDROADKM ------UNPAVEDROADKM ---(per thousandsquarekm)
(per thousandsquare nm)
LENGTH
GOOD
FAIR LENGTH
GOOD
FAIR
(km)
(km)
(km)
(km)
(km)
(km)
3.3
0.2
15.1
3.8
43.1
18.6
6.4
1.2
16.0
12.4
17.7
8.9
5.5
46.4
2.0
36.1
7.6
37.0
4.3
7.3
2.2
10.6
18.7
37.3
18.1
38.0
1.5
0.0
5.9
1.0
6.5
10.4
0.8
0.5
4.3
0.0
7.1
5.0
1.3
10.2
1.3
20.9
0.8
24.8
2.2
2.9
1.3
6.6
9.5
15.3
1.8
7.6
1.4
0.0
3.9
1.1
17.3
7.1
4.8
0.3
8.0
6.2
6.2
2.4
2.7
21.4
0.6
9.0
4.8
1.8
1.4
2.2
0.5
1.0
7.1
22.0
12.7
17.1
7.8
5.4
58.1
20.7
n.a.
62.5
10.7
24.8
23.0
n.a.
n.a.
9.0
26.4
72.7
7.9
110.3
18.5
27.2
7.3
20.1
3.1
50.5
32.7
182.9
n.a.
n.a.
3.7
n.a.
3.5
2.1
n.a.
5.0
0.6
10.9
6.9
n.a.
n.a.
2.4
n.a.
23.3
1.5
22.1
0.0
n.a.
0.3
6.0
0.7
4.0
16.7
34.9
n.a.
n.a.
2.4
n.a.
3.5
6.2
n.a.
47.5
4.7
7.2
9.6
n.a.
n.a.
3.2
21.1
28.4
1.0
62.9
26.2
2.7
0.7
7.0
0.9
18.7
12.5
84.5
n.a.
n.a.
ANNEX2 (continued)
DATAON POPULATION,
GNP/CAPITA,
AREA,PAVEDAND UNPAVEDROADS
-----
Countries
Pakistan
Comoros
Kenya
Togo
CA.R.
Haiti
Benin
Ghana
Lesotho
SriLanka
Guinea
Indonesia
Uberia
Mauritania
Sudan
Yemen
Senegal
Zimbabwe
Philippines
Morocco
Honduras
Svwaziland
PapuaNewGuinea
Egypt
Coted'lvoire
Thailand
Population
(million)
106.3
0.4
22.4
3.4
2.9
6.3
4.4
14.0
1.7
16.6
5.4
175.0
2.4
1.9
23.8
10.2
7.0
9.3
53.4
21.4
4.2
0.7
3.4
45.9
11.2
50.0
PAVEDROADKM--UNPAVED
ROADKM---(perthousandsquarekm)
(perthousandsquarekm)
GNP/
AREA LENGTH GOOD
FAIR LENGTH
GOOD
FAIR
Capita (thou.km2)
(km)
(km)
(km)
(km)
(km)
(kn)
350
370
370
370
380
380
390
400
420
420
430
440
450
480
480
522
650
650
660
670
700
700
710
720
770
860
796
2
583
57
623
28
113
239
30
66
246
1,905
111
1,031
2,506
625
197
391
300
447
112
17
463
1,001
323
513
30.3
246.5
10.8
26.3
0.7
21.6
9.2
25.1
20.1
134.8
5.3
14.5
5.0
1.5
0.9
15.5
19.2
32.6
47.3
33.1
14.4
40.5
1.6
15.1
12.3
54.6
5.5
105.5
3.4
10.5
0.2
0.0
2.4
7.0
10.7
13.5
2.6
4.3
4.3
0.8
0.3
6.0
5.4
8.8
14.7
6.6
7.2
14.2
0.5
5.9
9.2
27.3
15.2
130.0
5.6
5.8
0.2
21.6
4.6
11.2
5.8
40.5
0.0
4.3
0.6
0.4
0.4
8.6
6.1
1.0
25.5
14.6
6.2
14.2
0.7
5.0
3.1
16.4
n.a.
222.5
64.6
23.1
14.2
39.8
21.2
35.3
66.2
n.a.
23.2
n.a.
30.5
0.6
2.3
n.a.
33.0
58.7
n.a.
n.a.
n.a.
121.6
n.a.
n.a.
34.1
n.a.
n.a.
55.5
42.6
4.6
9.7
3.6
2.3
6.0
10.6
n.a.
n.a.
n.a.
4.6
0.1
0.5
n.a.
2.3
29.3
n a.
n.a.
n.a.
73.0
n.a.
n.a.
11.6
n.a.
n.a.
85.0
9.7
2.3
2.3
10.8
7.4
14.1
37.7
n.a.
na.
n.a.
22.9
0.2
0.5
n.a.
6.9
17.6
n.a.
n.a.
na.
45.0
n.a.
na.
22.1
n.a.
ANNEX2 (continued)
DATAON POPULATION,
GNPICAPITA,
AREA,PAVEDAND UNPAVED
ROADS
PAVEDROADKM------UNPAVED
ROADKM---(perthousandsquarekm)
(perthousandsquarekm)
GNP/
AREA LENGTH GOOD
FAIR LENGTH
GOOD
FAIR
Capita (thou.km2)
(km)
(km)
(km)
(km)
(km)
(km)
-----
Countries
Congo
DominicanRep.
Peru
Botswana
Cameroon
Bolivia
Belize
Ecuador
Jamaica
Guatemala
CostaRica
Tunisia
Turkey
Colombia
SyrianArab.Rep.
Chile
Brazil
Mauritius
Poland
Portugal
Panama
Mexico
Malaysia
Yugoslavia
Argentina
Romania
Populaffon
(million)
2.1
6.1
18.2
1.2
11.2
6.2
0.2
9.1
2.2
7.7
2.5
7.0
53.8
28.4
11.6
11.8
132.6
1.1
37.9
10.2
2.1
76.8
16.9
23.0
30.1
22.7
910
970
1,000
1,010
1,010
1,099
1,150
1,150
1,150
1,160
1,190
1,270
1,280
1,390
1,680
1,700
1,720
1,800
1,860
1,970
1,980
2,040
2,110
2,120
2,230
2,290
342
49
1,285
582
475
1,099
23
284
11
109
51
164
767
1,139
185
757
8,512
2
313
92
77
1,958
330
256
2,767
238
3.6
50.7
5.6
4.1
6.8
1.2
19.2
11.9
396.8
27.9
59.7
55.7
59.3
8.4
123.6
12.7
11.9
830.0
702.8
194.6
40.2
33.1
98.5
143.1
9.8
58.8
1.8
26.4
1.3
3.8
2.6
0.3
6.1
6.3
39.7
2.0
13.1
30.7
n.a.
3.5
na.
5.3
3.6
788.5
n.a.
97.3
14.5
28.1
n.a.
42.9
3.4
40.6
0.4
5.1
1.3
0.2
1.8
0.6
9.2
2.3
289.7
14.0
17.3
20.1
n.a.
3.1
n.a.
6.5
5.0
41.5
n.a.
58.4
21.7
3.3
n.a.
58.7
2.1
12.4
27.9
2.8
n.a.
10.0
61.9
8.3
71.7
17.0
363.6
5.5
74.6
n.a.
360.0
14.0
32.9
30.4
n.a.
70.0
449.4
n.a.
53.0
n.a.
23.1
n.a.
n.a.
n.a.
10.6
0.4
n.a.
4.5
9.9
1.7
11.5
11A
na.
1.6
13A
na.
n.a.
5.1
na.
2.7
n.a.
63.0
n.a.
n.a.
3.2
n.a.
n.a.
n.a.
n.a.
n.a.
7.5
0.8
na.
1s
35.3
3.0
54.5
3.7
214.5
1.0
45.5
na.
n.a.
6.9
na.
23.4
na.
3.5
n.a.
n.a.
34.4
n.a.
n.a.
n.a.
n.a.
n.a.
ANNEX2 (continued)
DATA ON POPULATION,GNP/CAPITA,AREA, PAVED AND UNPAVEDROADS
~~~~~ PAVEDROADKM ---Countries
Algeria
Uruguay
Gabon
Venezuela
Trinidadand Tobago
Cyprus
Korea,Rep.
SaudiArabia
Oman
Australia
UnitedKingdom
Italy
Belgium
Netherlands
Austria
France
Denmark
U.S.A.
Japan
Switzerland
Population
(million)
23.8
3.1
1.1
18.8
1.2
0.7
42.0
14.0
1.1
16.5
57.1
57.4
9.9
14.8
7.6
55.9
5.1
246.3
122.6
6.6
GNPI
AREA
Capita (thou. km2)
2,360
2,470
2,970
3,250
3,350
3,590
3,660
6,200
6,490
12,340
12,810
13,330
14,490
14,520
15,470
16,090
18,450
19,840
21,020
27,500
2,328
177
268
912
5
9
99
2,331
212
7,700
244
301
31
41
84
552
431
9,373
378
413
(perthousandsquarekm)
LENGTH
GOOD
FAIR
(km)
(km)
(km)
14.0
36.4
2.6
207.4
416.0
326.4
100.3
14.5
15.5
55.2
1443.8
1002.8
3973.7
2446.6
1275.0
1458.5
164.0
372.4
1948.6
172.0
Sources:WorldDevelopmentReport,Road Deteriorationin DevelopingCountries,
InternationalRoad Federation,and StaffAppraisalReports.
5.6
9.5
0.8
83.0
299.6
124.0
70.2
n.a.
10.2
46.9
1227.2
852.4
3377.7
2079.6
1083.7
1239.7
139.4
316.6
1656.3
146.2
4.5
21.5
0.8
78.8
79.0
124.0
25.1
n.a.
3.1
8.3
216.6
150.4
596.1
367.0
191.3
218.8
24.6
55.9
292.3
25.8
----
UNPAVEDROADKM---(per thousandsquare kin)
LENGTH
GOOD
FAIR
(km)
(kin)
(kmn)
n.a.
16.0
17.2
26.8
n.a.
n.a.
n.a.
24.7
n.a.
55.6
n.a.
n.a.
165.6
333.6
n.a.
n.a.
n.a.
292.6
972.8
n.a.
na.
7.4
5.5
6.2
n.a.
n.a.
na.
ria.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
na.
n.a.
n.a.
n.a.
7.5
5.1
16.1
n.a.
n.a.
n.a.
n a.
n a.
n.a.
n.a.
n.a.
n a.
n a.
na.
na.
n.a.
n a.
na.
n.a.
ANNEX
TIME SERIESDATA FOR
USA 1950 TO 1988
Year
Popn.
In m
1950
1951
1952
1953
1954
1955
1956
1957
1958
1969
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
152.3
154.9
157.6
160.1
163.0
165.9
168.9
172.0
174.9
177.8
180.7
183.7
186.5
189.2
191.9
194.3
196.6
198.7
200.7
202.7
205.1
207.7
209.9
211.9
213.9
216.0
218.0
220.2
222.6
225.1
227.8
230.1
232.5
234.8
237.0
239.3
241.6
243.9
246.3
GNP/Capita
1982dollars
7,903
8,168
8,270
8,524
8,280
8,942
9,128
9,130
8,730
9,304
9,402
9,431
9,854
10,076
10,447
11,160
11,748
11,952
12,392
12,511
12,263
12,525
13,339
13,884
13,403
13,159
13,717
14,228
14,646
14,523
14,043
14,020
13,617
13,966
14,774
15,122
15,389
15,800
16,339
Source:
a) HighwayStaftstics,DOT
b) StatisticalAbstracts
35
PavedRd/
mil. inhab
PavedRd/
thou. km2
8,240
8,455
8,729
9,236
9,417
9,660
9,926
10,150
10,469
10,760
10,952
11,150
11,345
11,625
11,797
12,049
12,113
12,308
12,619
12,843
13,007
13,131
13,323
13,372
13,675
13,818
13,927
14,395
14,529
14,196
14,409
14,475
13,910
13,774
14,529
14,187
14,34S
14,342
14,326
134
140
147
158
164
171
179
188
195
204
211
219
226
235
242
250
254
261
270
278
285
291
298
302
312
318
324
338
345
341
350
355
345
345
367
362
370
373
376
Policy Research Working Paper Series
Title
Author
Date
Contact
for paper
WPS906 Bulgaria'sEvolvingLegalFramework CherylW. Gray
for PrivateSectorDevelopment
Peter lanachkov
May 1992
CECSE
37188
WPS907 InstitutionalReformin Emerging
Securities Markets
May 1992
Z. Seguis
37664
WPS908 Tax Incentives,Market Power,and DagmarRajagopal
CorporateInvestment:A Rational
AnwarShah
ExpectationsModelAppliedto Pakistani
and TurkishIndustries
May 1992
C. Jones
37669
WPS909 ParallelMarkets,the Foreign
ExchangeAuction, and Exchange
Rate Unificationin Zambia
May 1992
V. Barthelmes
39175
WPS910 Policy Issuesin FinancialRegulation Dimitri Vittas
May 1992
W. Pitayatonakarn
37666
WPS911 Does ExchangeRateVolatilityHinder YingQian
ExportGrowth?Additional'vidence PanosVarangis
May 1992
D. Gustafson
33714
WPS912 Understandingthe InvestmentCycle AndresSolimano
in AdjustmentPrograms:Evidencefrom
ReformingEconomies
May 1992
E. Khine
39361
WPS913 The Women'sDevelopmentProgram MaitreyiDas
in Rajasthan:A CaseStudy in Group
Formationfor Women'sDevelopment
May 1992
L. Bennett
82772
WPS914 HealthPersonnelDevelopmentin
Sub-SaharanAfrica
J. Patrick Vaughan
May 1992
0. Nadora
31091
WPS915 Trade Policyand ExchangeRate
Issuesin the FormerSovietUnion
W. MaxCorden
May 1992
CECTP
37947
WPS916 Measuringthe Risk of Defaultin Six
HighlyIndebtedCountries
Marc Chesney
Jacques Morisset
June 1992
S. King-Watson
31047
WPS917 CreditorCountryRegulationsand
Asli DemirgOg-Kunt
CommercialBankLendingto Developing
Countries
June 1992
K. Waelti
37664
WPS918 Tax Evasionand Tax Reformin a
Low-incomeEconomy:General
EquilibriumEstimatesfor
Madagascar
Jaimede Melo
David Roland-Holst
MonaHaddad
June 1992
D. Ballantyne
37947
WPS919 Fiscaland Quasi-FiscalDeficits,
Nominaland Real:Measurement
and Policy Issues
RobertodeRezende
Rocha
FemandoSaldanha
June 1992
L. Ly
37352
RobertPardy
JanineAron
IbrahimA. Elbadawi
Policy Research Working Paper Series
Title
and Point
WPS920 EconomicIncentives
Choiceof
SourceEmissions:
Modeling
Platform
Contact
for paper
Author
Date
Raymond
J. Kopp
June1992
C. Jones
37754
June1992
M.Laygo
31261
andEconomic CesarOueiroz
WPS921 RoadInfrastructure
SurhidGautam
SomeDiagnostic
Development:
Indicators