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