1

April 1996 n Research Report 545

R E S E A R C HR E P O R T

Michigan AgriculturalExperiment StationMichigan State University

Impacts of Urbanization onAgricultural Sustainabilityand Rural Life in West Java,Indonesia

2

Impacts of Urbanization onAgricultural Sustainability and Rural Life

in West Java, Indonesia*

By:

Joyo Winoto and Gerhardus Schultink1

Department of Resource DevelopmentMichigan State University

East Lansing, MI 48824-1222, USA

3

* Research supported, in part, by funding and support from the Agricultural ExperimentStation, the Office of International Studies and Programs, and the Department of ResourceDevelopment at Michigan State University; the Research Institute and the Department of SoilSciences at Bogor Agricultural University; the Indonesian Center for Agro-Economics, theMinistry of Agriculture, the Indonesian Ministry of Education and Culture, the Central Bureauof Planning and Development, and the Asian Development Bank.

1 Dr. Winoto is the Director of Regional Resource and Environmental Development, Institute for National Development Studies, Indonesia, and teaches and conducts research at Bogor Agricultural University. Dr. Schultink is a Professor of International Resource Development at Michigan State University.

AAcckknnoowwlleeddggmmeennttssThe authors wish to thank the various institutions and their support staff members fortheir funding and logistical support. The views expressed in this paper are solely thoseof the authors and do not necessarily represent those of the national institutions orinternational funding agencies.

4

populations and to decrease rural-urban disparities in the near term,the government must change devel-opment policies and investmentstrategies targeting primarily non-agricultural sectors and urban areas.This may be accomplished by adiversion of regional investments topromote the establishment of gener-ative, secondary cities in the region— so-called “growth centers” —complemented with the develop-ment of a subregional infrastructure,effective regional planning andimprovement of rural service func-tions. In the longer term, the govern-ment must also provide greater localautonomy, at least, by fostering thedevelopment of local land use plansand effective legal controls. This canbe supplemented by incorporatingpolicy measures providingincentives for rural populationsengaged in agriculture, such as cre-ating transferable land developmentrights, preserving prime agriculturallands, open space conservation ease-ments, agricultural tax abatements,targeted subsidies in agriculturalinputs and marketing, and publicinvestments in the rural physicaland service infrastructure.

density as the differentiating criteri-on, the rate of regional urbanizationwas found higher in the “non-rural”areas, indicating a cumulative causa-tion process or self-reinforcingurbanization. Principal urbanizationimpacts include increases in agricul-tural land parcelization, landlessfarming households, absentee agri-cultural landownership and conver-sion of agricultural land to non-agri-cultural uses. Urbanization has alsotransformed self-employed farmersinto agricultural laborers and farmsto non-farm, small-scale enterprises.

Existing regional developmentstrategies have increased rural-urbandisparities, expressed as a compara-tive lack of economic progress andthe availability of and access to pub-lic facilities in rural areas.Urbanization has had a cumulativenegative effect on the viability of theregional food production system,specifically its important role as themajor contributor to the nationalpolicy of food (rice) self-sufficiency.Moreover, the welfare of the ruralpopulation engaged in theagricultural sector is decreasing.

To sustain regional agriculture, toovercome the negative impacts ofurbanization on the welfare of rural

AbstractCurrent Indonesian development

policies, including the regionaldevelopment strategy for theNorthern Coastal Region (NCR) ofWest Java, favor urban development.This research addresses the effects ofurbanization on the sustainability oftraditional agriculture—irrigatedrice production in a regionconsidered the “rice bowl” ofIndonesia. In addition, analysis isconducted on the effects of urbaniza-tion on the quality of life of the ruralpopulation engaged in agriculture,the regional urbanization processand its causal factors, its relationshipwith regional development and itsimpacts on agriculturaldevelopment. The goal is to identifypolicy interventions needed to sus-tain a viable regional agriculturalsector. The study employed threesets of data: primary data collectedat the village level for the period1982-1992, secondary data derivedfrom the censuses of 1980 and 1990,and secondary data provided by theProject of Land Conversion in Java.

Urbanization in the region iscaused mostly by economic stress inrural agricultural areas rather thanincreased economic opportunities inthe urban areas. Using population

5

IntroductionIndonesia’s population of 204 mil-

lion (World Handbook, 1995) isgrowing at about 1.6 percent peryear, with most of the increase occur-ring in urban areas. Expectations arethat almost 40 million residents willbe added to Indonesia’s cities in thefinal two decades of this century(Hamer, Steer and Williams, 1986).The most recent census indicates that62 percent (more then 126 million) ofIndonesia’s population is concentrat-ed on the island of Java2 (Moochtar,1992; Soegijoko, 1992; Soemarwoto,1992; Firman, 1992; Hill, 1992a; Hill,1992b; Soemarwoto et al., 1991;Hardjono, 1983), which accounts foronly 7 percent of Indonesia’s territo-ry (Poot, Kuyvenhoven and Jansen,1990; Hill, 1992a). Java, with anurban population of 23 million peo-ple in 1980, had about 70 percent ofIndonesia’s total urban population.The World Bank estimates that theurban population in Java alone willrise to 41 million by the year 20003,with Jakarta—the capital city ofIndonesia— having 26.46 percent,followed by West Java Province with25.17 percent. Outside Jakarta, mostof the urban population (34.24percent) is concentrated in West Java(Hamer et al., 1986), specifically theNorthern Coastal Region (Hill,1992b; Soegijoko, 1992; Firman,1992). Here, economic growth of theindustrial, trade and service sectorsacts as the driving force of urbaniza-tion. Firman (1992) concludes thaturban population growth in Java isevident not only in the big cities but

also in their peripheries and in theareas close to the regional arterialroads connecting the larger cities. Inother words, urbanization occurs inthe corridors joining large cities,such as the area between Jakarta andCirebon, usually referred to as JalurPantai Utara Jawa Barat or NorthernCoastal Region (NCR) of West Java.

Firman’s finding confirmsMcGee’s assertion that rapid urbangrowth in Asia is accompanied bythe emergence of peri-urban regionsand urbanized corridors joining bigcities4 (McGee, 1991). McGee (1987and 1990, cited in Firman, 1992, p.96) identifies six characteristics ofsuch urbanized corridors in Asia: (i)a very high population density; (ii)generally but not exclusively wet-rice regions with very smalllandholdings; (iii) enveloping bigcities in the regions; (iv) growth ofdiverse, non-agricultural activities;(v) considerable interaction betweenrural and urban activities; and (vi)an intense land use mix.

The study of the urbanizationprocess in such a corridor region is arelatively new phenomenon in theliterature of urbanization, whichusually focuses on the dynamics ofurban centers with an analytical con-text fitting developed nations ratherthan developing nations5 (Bhadraand Brandao, 1993; Datta, 1990).Although several studies have ana-lyzed this type of region, it is treatedas a small urban center withfunctions similar to those of largeurban centers6 (Hardoy and

Satterthwaite, 1986; Kabwegyere,1979; Southall, 1979; Rondinelli andRuddle, 1976; Hamer et al., 1986;Armstrong, Warwick and McGee,1985; Bhooshan and Misra, 1980;Honjo, 1981; McAndrew, 1990).Moreover, studies on the impacts ofurbanization on agricultural devel-opment in peri-urban and corridor-regions are rarely conducted indeveloping nations (Bhadra andBrandao, 1993, and Firman, 1992).Consequently, the nature ofurbanization and its impacts on therural service infrastructure and onagricultural development in suchurbanized corridors are not fullyunderstood.

Because regional growth compli-cates and modifies problems ofregional and agriculturaldevelopment, studies are needed toanalyze structural changes in urban-ized corridors to enhancedevelopment strategies and theirimplementation in the form of com-prehensive regional plans. Theseinclude studies to analyze: (i)impacts of changes on regionaldevelopment at the aggregate level(e.g., regional income, agriculturalsector performance and regionalinput-output accounts); (ii) impactsof urbanization on agriculturaldevelopment measured in basicsocioeconomic indicators, includingincome, employment and quality-of-life measures; (iii) changes in landuse patterns and land prices in thearea immediately surrounding maincities and in rural areas; (iv) the

2 Indonesia is a country composed of more than 13,700 islands, including the five biggest islands of Kalimantan, Irian Jaya, Sumatra, Sulawesi and Java.3 This estimate assumes that interprovincial and rural-urban migration patterns are unchanged from 1975-1980 rates. This assumption will

underestimate the real urban population growth because the dynamics of rural-urban migration as a result of infrastructure development in the past 15 years and the dynamics of economic development in the areas around big cities (Firman, 1992, p. 101).

4 McGee, in his article, Urbanisasi or Kotadesasi?: Evolving Patterns of Urbanization in Asia, 1989, called the development of urbanized regions such as the NCR a process of kotadesasi. “Kotadesasi is a coined word [in Indonesian or Bahasa Indonesia] that joins kota (town) and desa (village) to make up a word which carries the concept of urban and rural activity occurring in the same geographic territory” (McGee, 1989, pp. 93-94).

5 Bhadra and Brandao (1993, p. 2) assert: “The literature analyzing urban development is large. Over the last four decades, researchers have dealt with numerous related issues, such as explanations for the existence and growth of cities, the relationship between city growth and industrial growth, allocation and efficient use of urban sector-specific resources, availability and proper management of those resources and policies concerning balancedurban development.... Surprisingly, the literature has given much less attention to the implications of urban development planning and the process of land allocation between rural and urban uses, or, equivalently, between agriculture and non-agriculture. Moreover, most of the studies focus on the developed countries and in particular, on the United States.” Costa et al. (1989, p. 3) state: “While Asian urbanization is similar in many respects to Western urbanization, it is quite different as well.”

6 In their article, Why Small and Intermediate Urban Centers, Hardoy and Satterthwaite (1986, p. 6), for example, state: “...that it is small or intermediateurban centers which are the urban centers with which most rural people and rural enterprises interact. Yet the role that such centers can play in supporting social and economic development within rural areas...is rarely given sufficient attention.”

6

extent of agricultural land use con-version and its impacts on the wel-fare of the households and on agri-cultural sustainability; (v) changes inlinkages with metropolitan centers;(vi) changes in socioeconomic condi-tions in the region and their impactson migration—patterns of migrationwithin the region and migration tothe metropolitan centers; (vii)changes in rural and urban employ-ment; (viii) changes in the settlementpatterns; (ix) changes inenvironmental quality; (x) socioeco-nomic disparities between rural andurban areas within the urbanizedcorridors; and (xi) changes in rural-urban linkages in the area and inurbanized area-metropolitan centerlinkages.

This study analyzes some aspectsof the changes of the corridor region,especially impacts of regional urban-ization on agricultural development.The primary focus is on the corridorregion joining large cities subject tourbanization and with regional char-acteristics as those described byMcGee, as represented by theNorthern Coastal Region (NCR) ofWest Java, connecting Jakarta andCirebon. The NCR consists of fourkabupatens7: Bekasi, Karawang,Subang and Indramayu. This is boththe most urbanized region in Java(Firman, 1992) and the center of agri-cultural development, especially forwet-rice production8. As such, it istraditionally considered the ricebowl of Indonesia (Soemarwoto,1992).

Research ContextAbout 21 percent of Indonesia’s

GDP is contributed by outputs fromagriculture, forestry and fisheries,which involve slightly more then 50percent of the national labor force(World Handbook, 1995). Of the total

agricultural commodities, especiallyrice; subsidies in the developmentand rehabilitation of irrigated pad-dylands; provision of agriculturalinputs; and a buffer system to antici-pate price fluctuation in agriculturalcommodities (Nasoetion, 1994).

Most importantly, the GOIrestricts conversion of agriculturalland (especially irrigatedpaddyland) into non-agriculturaluses. This reflects the realization thaturbanization in this region has creat-ed development pressure on agricul-tural land. Also, through regionaldevelopment planning, which wasintroduced by the local governmentof West Java in 1974 (Government ofWest Java, 1990), the GOI restrictsagricultural land conversion(Presidential Decree No. 55, 1993,adopted into the Governor of WestJava Decision No. 593.05/SK.1785-Pem.Um/1993). This decree clearlystates that land use and land alloca-tion for development should followregional development planning(II.4.1) guidelines, which aredesigned to prevent the regionalconversion of agricultural land intonon-agricultural uses.

The development of the NCR can-not be separated from the develop-ment of Jakarta and West Java as awhole. According to Hill (1992b),industrial development — whichwas viewed as a necessary conditionfor economic and regional develop-ment in Indonesia during the pastthree decades — favored the regionbecause of its infrastructure andproximity to the seat of government.This notion implies thatdevelopment and investments favormetropolitan areas, which becomecenters of political power and com-mercial activity (Honjo, 1981a;Coates, Johnston and Knox, 1977; El-Shakhs, 1976). Furthermore,Mehretu (1989) and Mehretu, Wittick

land surface, 8 percent is arable land,of which about 40 percent is devotedto food production, with rice as thebasic staple crop. The NCR containsthe most productive agriculture, thebest irrigation systems (Soemarwotoet al., 1991) and the most advancedagricultural developmentinstitutions in Indonesia (Nasoetion,1994). Its development as an agricul-tural production center dates back tothe Dutch colonial period, when thedevelopment of the irrigationsystems started. Following the inde-pendence of Indonesia in 1945, theregion was further developedthrough the rehabilitation andexpansion of the irrigation systems.In the mid-1960s, the green revolu-tion9 in Indonesia—pioneered byBogor Agricultural University—orig-inated here. The policies andprograms implemented by theGovernment of Indonesia (GOI)included improvements in the irriga-tion infrastructure, continued invest-ment in the agricultural researchcapacity, the intensification of agri-cultural inputs, the introduction ofhigh-yielding varieties, continuedprovision of highly subsidizedinputs such as fertilizers and pesti-cides, and the development of theagricultural extension service. Allthese efforts have made the regionone of the centers of agriculturaldevelopment contributing toIndonesia’s self-sufficiency in rice in1985 (Erwidodo, 1990).

Besides agricultural productionsubsidies, the GOI also introducedseveral policies that directly or indi-rectly affect agriculturaldevelopment as well as urbanizationin this region.

At the macro level, GOI policy isto preserve self-sufficiency in foodproduction, especially rice.Consequently, the government main-tains policies of price controls on

7 Kabupaten is an administrative unit under province consisting of several kecamatan. In turn, one kecamatan consists of several desa, the smallest administrative unit equivalent to a village in the U.S. system.

8 ”At least half of West Java’s population occupies the Northern Coastal Region, which represents less than one-third of the total land area of the province”(Hehanussa and Hehuwat, 1979). The NCR is an alluvial lowland. It consists largely of alluvial river deposits and lahars [volcanic flow deposits] (Ongkosongo, 1979). Average annual rainfall is about 1880 mm (Ongkosongo, 1979), with average temperatures between 27.1o and 29.7o C (Ilahude, 1979).

9 The green revolution is a term used for rapid increases in rice yields in Indonesia brought about by improved varieties, irrigation and agricultural institutions with the expanded use of fertilizers and other chemical inputs (Andersen and Hazell, 1985).

7

and Pigozzi (1983) state that regionaleconomic and social development inmost LDCs have been characterizedby centripetal forces of concentrationand agglomeration of public as wellas private investments favoring rela-tively more developed regions.

In Indonesia, city residents havereceived priority access to central-ly funded, and heavily subsidized,public services, while the muchlarger and poorer population ofthe countryside has failed to keeppace. Metropolitan centers, inturn, have benefited most of all(Hamer et al., 1986, p. 9).

More recently, the GOIencouraged and subsidizedmedium- and large-scale industriesin this region (Hill, 1992b). Hill’sstudy shows that Java dominatesIndonesia’s non-oil manufacturingeconomy, generating almost threequarters of value-added products bymedium and large firms. West Javaalone contributes over one-third ofthis total, and its industrial outputexceeds that of the outer islands.Notably, most of the medium- andlarge-scale manufacturing industriesin this region are developed alongthe NCR, where the primeagricultural land is located.Moreover, the combined Jakarta-West Java value-added outputaccounts for more than 45 percent ofthe national total. This reflects thevery high degree of spatial concen-tration of industrial development.

Industrial development, with itsbackward and forward linkages, hasattracted more people to the NCR. Inother words, concentrated industrialdevelopment has induced urbaniza-tion in this region because of thetransfer of resources fromagriculture to non-agriculture andfrom rural to more urbanized areas.This transfer includes skilled andunskilled labor, capital and landresources (Bhadra and Brandao,

1993), and service provision (Unwin,1989).

The increased transfer of resourcesover time is determined mainly bythe supply of and demand forresources in the growing regionaleconomy. Bhadra and Brandao (1993,p. 3) describe such a transfer processas follows:

From the demand side, the incen-tive for resource transfer arises, inpart, from larger income sharesspent on industrial products andservices relative to that on agricul-tural goods.... The supply sideeffects come from faster technicalprogress in industry than in agri-culture...and from significant scaleeconomies in urban productionprocess.

More importantly, the resourcetransfers “are not only symptoms ofthe ‘development process’ but arethemselves active features in thetransformation of rural and urbanplaces” (Gould, 1982, p., cited inUnwin, 1989, p. 13).

The modernization theory ofurbanization, furthermore, explainsthat urbanization occurs when theregion faces structural economictransformation (Choe, 1981, p. 101),characterized by the increased rela-tive share of the industrial and ser-vice sectors and the decreased rela-tive share of the agricultural sectorin the regional economy (Timmer,1990). Moreover, “...the rate ofurbanization or equivalently the rateof labor redistribution to the urban-industrial sector is largely dictatedby industrialization...” (Choe, 1981,pp. 100-101). The implicitassumption behind this line of think-ing is that urbanization is a naturaland inevitable consequence of indus-trialization and modernization(Sharbatoghlie, 1991, p. 11).

Whatever the reason behindurbanization occurring in the NCR,the process puts pressure on the

agricultural resource base, especiallyon prime agricultural land (Schmid,1968) destined to support and main-tain national food self-sufficiency(Firman, 1992; Soegijoko, 1992;Soemarwoto, 1992; Soemarwoto etal., 1991).

Parallel to the described industrialdevelopment in the NCR, the WestJava regional development plandesigned in 1974 has acceleratedurban-biased development, concen-trating investments in the primarycities. In addition, urbandevelopment has acceleratedbecause West Java adopted function-al integration rather than territorialintegration10 as its regional planningstrategy with “growth centers”11 asthe principal element of regionaldevelopment policy.

At a broader level of regionaldevelopment (Jakarta-West Java),economic and social developmentwere to have focused on the bigcities, Jakarta and Cirebon, with thelatter to balance the development ofJakarta. At the kabupaten level,development efforts andinvestments were to have focused onthe primary city or capital city of thekabupaten. This regionaldevelopment strategy has put agri-cultural development in the NCR atrisk, even though it was designatedas the center of agricultural develop-ment with major government invest-ments in the agricultural sector(Nasoetion, 1994). Therefore, thedevelopment of Jakarta and Cirebon,as well as the development of prima-ry cities in each kabupaten, will putsome pressure on the agriculturalresource base. According toFriedman and Wolff (1992), this typeof policy in regional developmentwill wash out higher qualityresources in the rural agriculturalhinterlands. Accordingly, smallercities or centers of development willbe used as outposts by the urban

10 ”The territorial force derives from common bonds of social order forged by history within a given place. Functional ties are based on mutual self-interests. Given inequalities at the start, a functional order is always hierarchical, accumulating power at the top. Territorial relationships, on the other hand, though they will also be characterized by inequalities of power, are tempered by the mutual rights and obligations which the members of aterritorial group claim from each other” (Friedman and Weaver, 1979, p. 7).

11 Growth center is defined as “induced urbanization through a combination of direct public investments and capital subsidies to private enterprise” (Friedman and Weaver, 1979, p. 6).

8

ruling class and urban interest toachieve three related purposes: toextract a sizable surplus from therural economy, chiefly in the form ofprimary products through a processof unequal exchange; to expand themarket for goods and services pro-duced in urban centers; and toensure stability of the politicalsystem that maintains urban domi-nation over the rest of the region.

Macro and regional developmentpolicies implemented in the NCRseem to conflict with the GOI’sintention to maintain the NCR as the“rice bowl” of Indonesia. Even as theGOI has launched several policies toprotect the agricultural resource baseand to maintain agricultural produc-tion in the NCR, current macro andregional policies put agricultureunder pressure from urbanization.This is, in part, induced by regionalindustrial development as well asurban-biased development.Moreover, the NCR itself has alsofaced massive urbanization in thepast two decades (Firman, 1992;Soemarwoto, 1992).

Urbanization in the NCR, both asa corridor region connecting Jakarta-Cirebon and as a region designatedas the rice bowl of Indonesia, repre-sents a public policy challenge inIndonesian agricultural and regionaldevelopment. Assuming thatIndonesia will be able to maintainand improve food security in thelong run by expanding agriculturalproduction to the outer islands or byinternational trade, it may be expect-ed that the role of the NCR as therice bowl of Indonesia will be dimin-ished. Under this assumption, sacri-ficing agriculture to accelerateregional economic growth throughindustrialization might be feasible.Totally abandoning agriculturaldevelopment in this region, howev-er, will not be socially and economi-cally acceptable for three fundamen-tal reasons:

a. Economic (opportunity andsocial) cost: The government hasalready made large capital

between agricultural developmentand industrialization, and assertedthat agricultural stagnation, espe-cially food shortage, could chokeoff growth in the non-agriculturalsectors by affecting labor supply.

c. Agricultural sustainability:Pingali, Moya and Velasco’s (1990)study about the post-green revolu-tion in the Philippines, Thailandand Indonesia used West Java, a“rice bowl” province, as a sample.It concludes that: there is a trendtowards stagnation and/or adecline in irrigated rice yields, andthe rate of degradation of thepaddy environment is greater thanthe rate of growth in yield poten-tial. They also asserted that“...while national average yield [ofrice] continues to rise, [the] aver-age yield in the province has beenstagnant...” (Pingali et al., 1990, p.11). This can affect net farm house-hold income since the studyshows that the increased use ofmodern agricultural inputs doesnot result in an increase in produc-tivity. The decrease in householdnet income will discourage farm-ers from staying in agriculture. Inaddition, the increased rate ofdegradation of paddy rice produc-tion will reduce the long-termagricultural productive capacity ofthe region. These imply that theagricultural viability of the regionis at risk. Therefore, whether or not other

regions in Indonesia can increasetheir share in maintaining the nation-al food security, sustaining a viableagricultural sector in this region is anational public policy imperative —short- and long-term — especiallyconsidering development objectivespertaining to improving the welfareof rural populations.

This study starts with theassertion that urbanization in theNCR and Jakarta-West Java regionalplanning have negatively affectedagricultural development in theNCR and assesses causalrelationships between urbanization

investments in agricultural devel-opment in this region resulting inlarge opportunity costs if agricul-ture in this region is to besacrificed (Soemarwoto et al.,1991). Most of the agriculturalland in this area is irrigated pad-dyland, not dryland farming sys-tems, and it typically takes morethan 10 years to establish produc-tive paddyland. Establishmentcosts are in addition to the directexpenditures associated with thedevelopment of irrigationnetworks and relatedsocioeconomic costs (Winoto,1985). Even more important maybe the large social cost associatedwith uncontrolled urbanization.Although the agricultural sector’srelative share of the regional econ-omy has been declining over time,more than 60 percent of the laborforce in this region still derives itsincome from agriculture(Soemarwoto, 1992). The develop-ment of the urban-industrial sec-tor has not been able to absorb thelabor surplus from the agriculturalsector (Timmer, 1990).Consequently, the land-labor ratioand labor productivity per unit ofland in agricultural areas willdecline (Naylor, 1992) and incomeper capita will also decline12

(Eicher and Staatz, 1990). Furthermore, Erwidodo (1990)shows that this conversion processhas also resulted in fragmentationor disaggregation of agriculturalland per household. The percent-age of households on farms of lessthan 0.5 ha was 45.7 percent in1973 and 63.3 percent in 1980. Thenumber of landless farmers hasconsistently increased at a higherrate than the number of landown-ers. The landless made up 3.2 per-cent of the total farm householdsin 1973 and 14.9 percent in 1980.

b. Labor supply constraints: Classicstudies by Jorgenson (1961), Ranisand Fei (1961), Enke (1962a; 1962b)and Ranis (1963; 1964) establishedthe theoretical relationship

12 Assuming that the increase in productivity from technological innovation is less than the increase in agricultural labor supply.

9

and agricultural development.Specifically, it describes the spatialeffects of existing regional develop-ment policies on the distribution ofbenefits of regional development,especially between rural and urbanareas of the region; analyzes thenature and process of urbanizationand determinant factors affectingurbanization; and describes effects ofurbanization on agricultural devel-opment, specifically the structure ofland tenure, land conversion and itsimpact on household income, andthe structure of rural employment.

Research ObjectivesThe general goal of this study was

to provide a better understanding ofthe urbanization process and its deter-minants, its relationships with region-al development and, most important-ly, its impacts on agricultural develop-ment in the NCR, an urbanizing corri-dor region between Jakarta andCirebon. Understanding this generalobjective provides a useful means toidentify policy interventions tosustain a viable agricultural sectorsubject to urbanization. The studyobjectives are to:1. Analyze urbanization in the NCR

and identify factors associatedwith urbanization, such aschanges in land tenure, changesin agricultural economy,demographic changes,educational development, devel-opment in public facilities, indus-trial development and proximityto primary cities.

2. Describe the impacts of urbaniza-tion on agriculturaldevelopment, especially landtenure — the structure oflandownership, the degree ofabsentee landownership in theagricultural sector, agriculturalprime land conversion and itsfollowing impact on agriculturalhouseholds’ income; the impactsof urbanization on the rural andregional structure ofemployment; and the impacts ofurbanization on rural industrial-ization.

urbanization as the increase in popu-lation density and an increase in het-erogeneity of a wide range of non-agricultural occupations. The litera-ture of urban geography providessome fundamental aspects of urban-ization that facilitate thedevelopment of indicators of urban-ization. Breese (1966, p. 3) definedurbanization as a “process of becom-ing urban, moving to cities, changingfrom agriculture to other pursuitscommon to cities, and correspondingchanging of behavior patterns.” Bydefinition, urbanization involves achange in economic orientation fromagricultural to non-agricultural activ-ities and a migration from rural tourban areas. Schmid (1968) gave as aspecific indicator of urbanization theincrease in land values in the rural-urban fringe.

Friedman (1966) gives two broaddefinitions of urbanization that areapplicable to urbanization in lesserdeveloped countries (LDCs). First,“urbanization commonly refers tothe concentration of formerlydispersed populations that areprimarily engaged in farming in asmall number of settlements whoseprincipal economic activities are inthe services, trades, andmanufactures.” The second meaning“refers to urban modes ofproduction, living, and thinking orig-inating in these centers and spread-ing from these to outlying towns andrural populations” (Friedman andWulff, 1982, p.1). These definitionsare modified by Williams, Brunn andDarden (1983, p.5) when they defineurbanization as “a process involvingtwo phases or aspects: (i) the move-ment of people from rural to urbanplaces where they engage primarilyin non-rural functions oroccupations; and (ii) the change inlife-style from rural to urban with itsassociated values, attitudes, andbehaviors. The important variables inthe former are population densityand economic functions; the impor-tant variables of the latter depend onsocial, psychological, and behavioralfactors. The two aspects are mutuallysupportive.”

3. Describe the spatial impacts ofurbanization on the distributionof the agricultural resource base,the distribution in thedevelopment of public facilities,and the disparities in economicdevelopment in rural and urbanareas within the region.

4. Describe the relationships amongurbanization, regional planningand agricultural development.

5. Examine the policy implicationsof the findings for sustainingagricultural and developmentplanning in the NCR

Urbanization andAgricultural Development:Key Concepts

Two important concepts employedin this study are urbanization and agri-cultural development. They reflect theliterature and theories of urbanizationand agricultural development, andthe uses of these theories to measurethe impacts of urbanization onagricultural development.

Urbanization is a term with differ-ent meanings to different scientificdisciplines and specialists (Firman,1992). For regional economists,urbanization might mean theprocess accompanying the structuraltransformation of the regional econ-omy (Reismann, 1964; McGee, 1971;Choe, 1981; Bhadra and Brandao,1993). For land resource developers,urbanization might mean conversionof agricultural lands to non-agricul-tural uses and an increase in realestate values (Schmid, 1968). For asociologist-demographer, the urban-ization process means the process ofmigration from rural to urban areaswith its sociological and spatial con-sequences (Hauser et al., 1985). Forpsychologists, urbanization mightmean the changing of personal ori-entation from a rural-traditional tourban-modern way of life (Helmerand Eddington, 1973).

The importance of the growth ofnon-agricultural activities in definingurbanization was also emphasizedby Sjoberg (1960) when he defined

10

Urbanization in Indonesia is usu-ally defined according to CentralBureau of Statistics’ (BPS, 1988) crite-ria13. As presented by Firman (1992,p. 97), an urban area is commonlydefined as an area that has: (i) a pop-ulation density of 5,000 persons persquare kilometer or above; (ii) 25percent or fewer agricultural house-holds; and (iii) eight or more typesof urban facilities. Murphey (1966)suggested that the study ofurbanization would be most useful ifthe researcher used the criteria ofurbanization adopted by the nationwhere the study was conducted. Hissuggestion was based on his experi-ence in Indonesia and China, wherereliable data for the study of urban-ization—such as conducted inMDCs—were not available.

This study follows the Friedman(1966) and Williams et al. (1983) defi-nition of urbanization so that a sin-gle indicator of urbanization can bedeveloped. In a more recent study ofurbanization, this line of thinking indefining urbanization is employedby Sharbatoglie (1991, p. 51) whenhe defines urbanization as “theprocess which results in an increas-ing number of people living in non-agricultural settlements.”

Thus, urbanization, in this study,is defined as a concentration of for-merly dispersed populations accom-panied by the changing occupationalorientation to non-rural or non-agri-cultural functions. Suchurbanization can be measured byusing at least three indicators —population density, land rent andeconomic functions. This study,however, will use the rate of changein population density to describe theurbanization process. The use of thissingle indicator as a measure ofregional urbanization in the studyarea is based on the simple fact thatdata on population density are avail-able at the desa (or village) level andmonitored by national census every10 years. In addition, it is also basedon the theoretical statementproposed by Adna Weber (cited in

agricultural development should be“directed to the well-being of peoplerather than on agricultural growthitself” (Cramer and Jensen, 1988, p.334). In formulating an operationaldefinition of AD, this last study usedan eclectic approach that allows theclassic theory of development, thetheory of sustainable development,the classic theory of agriculturaldevelopment and sustainableagricultural development to belinked.

The starting point in formulatingan operational definition ofagricultural development is to tracethe meaning of “development” itself.Although development is a verybroad concept, the literature of devel-opment has provided some importantcomponents of development that canbe applied in this study.

With the assumption that socialand political problems can beovercome if the national incomeincreases more than the populationgrowth, most development theoristshave equated development with eco-nomic development and economicdevelopment with economic growth(Eicher and Staatz, 1990).Consequently, development effortshave been directed mainly toincrease national income(Mabogunje, 1981; Santos, 1977; andDadzie, 1980). But the experience ofthe past decades has shown thatsocial problems and politicalupheavals have emerged in countriesat all stages of development, both incountries with rapidly rising percapita incomes and those with stag-nant economies (Seers, 1970).Therefore, the theory of developmentbecomes critically questioned bymost development theorists (Gillis etal., 1987; Little, 1982). According tosome of these development theorists,“...development is a normative con-cept which is almost a synonym forimprovement...” (Okun andRichardson, 1961; Seers, 1970;Colman and Nixson, 1978; Honjo;1980; Bhooshan and Misra, 1980;Bryant and White, 1982).

Berry, 1981, p. 27), “...urbanization isa process of populationconcentration. It proceeds in twoways: the multiplication of thepoints of concentration and theincreasing in size of individual con-centrations.... Just as long as citiesgrow in size or multiply in number,urbanization is taking place....Urbanization is a process of becom-ing. It implies a movement from astate of less concentration to a stateof more concentration.”

The selection of change in popula-tion density as the indicator ofregional urbanization in this studyfollows a classification of rural andurban areas based on populationdensity. This classification may seema rather simplistic measure because itdoes not reflect the dynamics of eco-nomic activities in both rural andurban areas. However, this measurecan be used as a surrogate of urbanphenomena by interpreting less pop-ulated areas as real rural areas andmore populated areas as areas withurban characteristics. Moreover, thismeasure is commonly used to classi-fy rural and urban areas (Firman,1992; Sharbatoglie, 1991; Hamer et al.,1986; Sutton, 1989; El-Bushra, 1989;Obudho and Waller, 1976; Nath,1989). The main difficulty in usingpopulation density as the indicator todifferentiate rural and urban areas isin determining the cutting point ofpopulation density representing thetransition between rural and urbanareas. Various studies have used dif-ferent density levels to differentiaterural and urban areas.

The definition of agricultural devel-opment (AD) adopted in this studygoes beyond Hayami and Ruttan’s(1985) concept that emphasizes thetotal production and productivity ofagriculture. It adopts the concept ofsustainable agricultural development(SAD) with its main focus on theresults of agricultural development— namely, to maintain and improvethe welfare and well-being of thepeople who earn their livelihood inagriculture. It is assumed that

13 In Indonesia, the differentiation between urban and rural areas is made at the village or desa level, so there are urban villages (desa kota) and rural villages (desa desa).

11

Two important concepts areembedded in the theory of develop-ment. The first is a concept of valuethat should be followed in the devel-opment efforts; the second is theconcept of improvement. In the valueconcept, Seers (1970) proposes thenecessary conditions for a universal-ly acceptable norm as the realizationof the human personality. The neces-sary conditions are enough food(physical necessity) and income tocover other basic needs such asclothing and shelter, and, moreimportantly, social justice, reflected,in part, in income distribution.Accordingly, the concept of“improvement” represented indevelopment is overcoming poverty,unemployment and inequality, aswell as fulfilling human potentialthrough freedom, education andpolitical independence. This theorymeasures development by usingbroader indicators than just theincrease of national income.Moreover, some indicators of devel-opment (e.g., basic needs, equity,poverty reduction) are incorporatedin the theory of sustainable develop-ment, although sustainable develop-ment is more than the theory ofdevelopment proposed by Seers(1970).

Sustainable development emphasizesmoral responsibility, which is rootedin the philosophy that the currentgenerations have a moral obligationtoward future generations (Mellert,1985) because: (i) future generationswill be essentially the same as thepresent generation; (ii) one is borninto a given generation by historicalaccident; (iii) our survival as aspecies is more important than ourindividual survival; and (iv) evenafter we die, the effect of our livescontinues. Our obligation is based onthe truth that we are more thanunique and separate individuals, liv-ing only the immediacy of the now.We are, rather, parts of a much largerwhole, one that transcends space andtime.

The generic definition of SD is“paths of human progress to meetthe needs of the present generation

without compromising the ability offuture generations to meet their ownneeds” (UNCED, 1987). This defini-tion has fundamental objectives —meeting current needs while consid-ering future generations’ needs —from which we can derive a range ofoperational objectives that cut acrossmost previous intellectual and politi-cal boundaries (Lele, 1991).

The concept of sustainable develop-ment originated in the context ofrenewable resources and has subse-quently been adopted as a broaderparadigm by the environmentalmovement. Agriculture, consequent-ly, as one of the foundations ofhuman society and as a major activi-ty at the human-environment inter-face, receives the most attention inthe operationalization of sustainabledevelopment (Lele, 1991). The con-cept of sustainable agriculture is stillevolving from the dialogue concern-ing the issue of agriculture and theenvironment (Ikerd, 1990). Douglas(1984) identifies three schools ofthought in sustainable agriculturaldevelopment (SAD).

The first concerns SAD as foodsufficiency — expand the supply offood with some combination of moreresources and greater productivecapacity. For this school of thought,the problem of sustaining theresource base or of honoring the cul-ture of agriculture seems less impor-tant than the challenge of keepingup with population growth. Ruttan(1990) supports this school ofthought. He believes thatsustainability is not enough, that “ifthe concept of sustainability is toserve as a guide to practice, it mustinclude the use of technology andpractices that both sustain andenhance productivity” (Ruttan, 1990,p. 402).

The second sees SAD primarily asan ecological question. It concernsthe ecological balance. For this schoolof thought, the agricultural systemthat needlessly depletes, pollutes ordisrupts the ecological balance ofnatural systems is unsustainable andshould be replaced by one that hon-ors the longer term biophysical

constraints of nature. Instead of tak-ing population as given, thisapproach tends to espouse policiesthat limit population to those levelsthat a finite physical environmentcan sustain.

The third is an alternative agricul-ture or radical agriculture. Its atten-tion focuses primarily on the effectsof various agricultural systems onthe social organization and cultureof rural life. According to DESFIL(no date), this school of thoughtresembles the ecologists in theirdesire to husband the permanentcarrying capacity of renewableresources, but it differs in its empha-sis on sustainable human communi-ties. Not only must human beingsestablish stewardship of the earth —they must also establish this sense intheir relations with each other, par-ticularly as it affects justice and par-ticipation.

According to Lockeretz (1986), thislast school of thought has broadergoals than the other two schools ofthought. Accordingly, it can be takento refer to agricultural systems thatshare certain broad goals: (i) to con-sider the structure of the agriculturalsector that favors small to moderate-sized farms, especially independentfamily farms worked by residentowner-operators; (ii) to reduce theties to the industrial economy byincreasing the self-sufficiency offarms; (iii) to strengthen local andregional agriculture to achieveregional food self-sufficiency; (iv) topreserve farmland and avoid furtherdisaggregation of farmland; (v) tobetter protect the agriculturalresource values; and (vi) to achievelong-term sustainability throughconservation of the finite agricultur-al resource base.

Across these three schools ofthought, Schultink (1992a, p. 206)defines SAD as “the developmentand management of naturalresources to ensure or enhance thelong-term productive capacity of theresource base and improve the long-term wealth and well-being derivedfrom alternative resource usesystems, with acceptable

environmental impacts.” Therefore,in SAD, policy and objectives shouldbe designed to accomplish one majorgoal, “the achievement of asustained flow of benefits whichenhance the quality-of-life of humanpopulations without reducing thelong-term productive capacity of ourresource base” (Schultink, 1992a, p.204).

For its policy purposes, the GOIhas defined SAD as “...developmentin which social benefits exceed socialcosts, considering a long time frameand special consideration for thepoor and vulnerable. It implies, forpresent and future generations, con-tinuous improvement in real percapita income and quality of life,continuous narrowing of income dis-tribution, elimination of physicalsuffering due to poverty....” (Tarrantet al., 1987).

In defining agricultural develop-ment, Stevens and Jabara (1988, pp.5-6) suggest that the following direc-tions are included in the frameworkof agricultural development. Theypropose that agriculturaldevelopment should be directed to:(i) increase real per capita income ofthe agricultural population; (ii)increase food security; (iii) reducethe number of people in poverty; (iv)achieve the least possible amount ofdisruption in cultural values, rurallife and employment; (v) achieve adesired level of equity; (vi) establisheconomic incentives that encourageincreased productivity, creativityand enterprise; and (vii) increaseindividual opportunities, economi-cally and politically.

Based on the previously discussedconcepts of agriculturaldevelopment, this study uses theoperational definition of agriculturaldevelopment in LDC settings as aneffort directed to maintain andimprove the agricultural support infra-structure, to promote a cost-effective

structure of agricultural landownership,to reduce the dependence on absenteelandownership with the objective tomaintain and improve food security ofrural populations, to enhance the welfareor quality of life of the rural populationemployed in agriculture, and to reducerural-urban disparities without reduc-ing the long-term productive capacity ofthe agricultural resource base.

The assumption embedded in thisdefinition — that improvement ofthe rural population’s food security,quality of life and rural-urban equitycan be fostered by maintaining andimproving the agricultural resourcebase, and strengthening the landtenure, such as by reducing thedependency on absentee landowners— is that this can be accomplishedby an improvement of regionaldevelopment strategies. Thisassumption may be supported byseveral researchers and developmentpractitioners. Smith (1982, p. 19)states that, “...ownership of wealth,in the form of money, land or otherassets, provides direct access tosome of the good things in life byvirtue of providing purchasingpower or social status.” Therefore,the agricultural household’s foodsecurity, welfare and quality of lifecan be attributed to the securedownership of agricultural lands asthe basis of the agricultural produc-tion system (Sen, 1990; Eicher andStaatz, 1990; Stevens and Jabara,1988; Chuta and Liedholm, 1990).

In this aspect, this study is limitedto analyzing the changes in the struc-ture of landownership, the changes inthe degree of absenteelandownership, the changes in agri-cultural prime land (or agriculturalinfrastructure, based on the WorldBank’s definition [World Bank, 1994])and its subsequent impact on agricul-tural household income, changes inthe structure of rural employmentand rural-urban disparities.

Data Collection andResearch Method

This study used three sets of data.The first set was primary datacollected through village surveys.These data represented a regionalsample with the desa (village) as theunit of analysis. Secondary datawere provided by the AgriculturalLand Conversion Project, based onan agricultural household survey ofirrigated land conversion in the past10 years. The third was 1980-90 cen-sus data.

The primary data, collected in1982 and 1992, updated informationfrom census data, collected every 10years. Using the village as a unit ofanalysis made it possible to differen-tiate desa kota (urban villages) anddesa desa (rural villages).

The information covered 1982-1992. The data include: 1. General information on the

villages such as village area(hectare [ha]), total paddyland14

(ha), agricultural dryland (ha),land used for housing (ha), andother lands (ha).

2. Village population — totalhouseholds and total populationas well as total numbers of menand women.

3. Population distribution based onsources of income15 such asfarmer (landowner),sharecropper, agricultural laborer,small-scale enterprise16,government official, private sector (entrepreneur)17, manufac-turing labor and informal sector18.

4. Migration, including total in-migration and total out-migration by sex.

5. Village distance to the cities andmarket (kilometer [km]) includ-ing the distance from the villageto Jakarta (km), to Cirebon (km),to the capital city of the

12

14 ”Paddyland” and “irrigated paddyland” are used interchangeably in this study.15 Based on the largest share to the total individual income and defined as more than 50 percent of the total income.16 Small-scale enterprise is defined as an activity related to subsistence that needs labor from outside the household. 17 Entrepreneurs (private sector) are defined as people who have their own businesses and use outside labor.18 Informal sector is defined as an activity that is not captured in the definition of small-scale enterprise, such as angkot driver (angkot is a small bus) or

ojek (a transportation system using motorcycles).

13

kabupaten (km), to the capitalcity of the kecamatan (km) and tothe local market used mostly bythe village population (km).

6. Infrastructure and social-economic conditions such as totalpaved road (km), total hardenedroad (km), number and types ofpublic transportation modes,number and types of economicinstitutions in the village such asbanks or cooperatives, numberand types of health care facilities,number and types of educationalfacilities, number and types ofsocial and religious facilities suchas youth centers or mosques,number of households with elec-tricity, number of householdswith televisions and number ofhouseholds with telephones.

7. Village industries, includingnumber of small scale-, medium-and large-scale industries19.

8. Paddyland conversion (ha); theprice of paddyland (Rp.,[Rupiah]) and the number ofland transactions in the village.

9. Average agricultural wage rate(Rp.) in the village, based onplowing and harvest periods.

10. Landownership of dryland andpaddyland, absentee landowner-ship of dryland and paddyland,and number of absenteelandlords.

11. Some qualitative data, such asvillage perception about theurbanization process in theregion or problems with compen-sation for land conversion.

The sampling procedureemployed follows Barber’s (1988)process, consisting of five steps:

1. Definition of population, i.e., acollection of all individualelements from which the sampleswill be collected. The unit ofanalysis of this study is the desaor village, and the populationconsidered in this study

represents all villages in theNCR.

2. Construction of a samplingframe or a population frame, i.e.,an ordered list of population ele-ments. The sampling frame orthe population frame was devel-oped using a list of villages and avillage-level map of eachkabupaten. In this study, the tar-get population was the same asthe sampled population20.

3. Selection of a sampling design,i.e., a procedure used to selectelements from the samplingframe for the sample. There arefour sampling designs commonlyemployed in the data collection:simple random sampling, strati-fied sampling, systematicsampling and cluster sampling(Barber, 1988; Williams, 1986;Berry and Baker, 1968).According to Barber (1988, p.216), in terms of efficiency,“Stratified samples are [the] bestand cluster samples are [the]worst. Random samples usuallylie somewhere between these twoextremes. That is, for a givenlevel of precision, stratified sam-ples require the fewestrespondents and cluster samplesrequire the most respondents....”

To optimize sampling efficiency(highest confidence level at agiven sampling size), the studyused a stratified, random samplein each kabupaten in the NCR asthe sampling design. Eachkabupaten in the NCR was strati-fied into four strata based on thevillage population density, whichreflected the degree ofurbanization. The first stratumwas villages with up to 750 peopleper square kilometer; the secondstratum was villages with popula-tion densities between 751 and1500 people per square kilometer;the third stratum was villageswith population densities between

1501 and 2250 people per squarekilometer; and the fourth stratumwas villages with population den-sities of more than 2250 people persquare kilometer. In each of thesestrata, the village samples wererandomly chosen. Total samples for this study were44 villages or desas. According tothe central limit theorem, samplesizes of > 30 can be expected torepresent sample characteristicson which basis inferences aboutthe population can be made withconfidence.

4. Specification of the informationto be collected.

5. Collection of the data. In thisstudy, data collection or datagathering was divided betweentwo groups. The first group cov-ered the kabupatens Bekasi andKarawang and was coordinatedby the Center of Agro-EconomicResearch (PAE). Within thisgroup and for each kabupaten,data were collected by three enu-merators from the office of PAEaccompanied by one kabupatencoordinator. The second group,which covered kabupatensSubang and Indramayu, wascoordinated by the ResearchInstitute, Bogor AgriculturalUniversity (IPB). Within thisgroup, data collection in bothkabupatens was conducted bythree enumerators from IPBaccompanied by onecoordinator,21 who was alsoresponsible for gatheringsecondary data in all four kabu-patens in the region. All enumer-ators were trained to ensure thatthey understood the questionsand concepts. Prior to the train-ing, the coordinators conductedseveral meetings to develop thequestionnaire and coordinate thesurvey.

Two sets of secondary data weregathered from several sources. The

19 Large-scale industry is defined as an industry with 100 or more employees; medium-scale industry is defined as an industry with 20 to 99 empoyees; and small-scale industry is defined as an industry with more than 5 to 19 employees.

20 “The target population is the set of all [elements] relevant to a particular study. The sampled population consists of all the [elements] listed in the sampling frame” (Barber, 1988, p. 206).

21 Dr. Winoto was the coordinator of this group.

14

first data set represented 1980 and1990. Census data collected at thedesa level contained informationrelated to population and villageeconomy. The census data coveredall 1099 villages in the NCR. The sec-ond data set related to the impacts ofagricultural land conversion onhousehold income. These data wereprovided by the Project ofAgricultural Land Conversion inJava, Center of Agro-EconomicResearch (PAE), Ministry ofAgriculture, which had conducted ahousehold survey in the NCR in1993. The respondents of that surveywere the households involved inland conversion during the past 10years. The total number ofrespondents was 100 households,consisting of 52 households involvedin land conversion and 48households not involved in landconversion. For the householdsinvolved in land conversion, income—both farm income and off-farmincome before and after land conver-sion — was recorded. The surveyused a random sampling techniquefor each group of agricultural house-holds.

Compilation andOrganization of Data

Primary data and secondary datafrom the 1980 and 1990 censuseswere tabulated, edited and compiledat Michigan State University (MSU);the secondary data related to house-hold income were compiled andorganized by the office of PAE,Bogor, Indonesia. The first two datasets were organized using the stan-dard geographic matrix recommend-ed by Berry (1964). This standardmatrix is widely used in regionaleconomic analysis. It eases theprocess of statistical analysis by con-sidering both the spatial and timedimensions. Most importantly, thisorganization of data was compatiblewith the structure of data in theSPSS, the analytical software used inthis study.

Data AnalysisTo address the research objectives,

the data were analyzed in thefollowing sequence :1. The study area was stratified into

two main strata, rural and urban,based on population density. Forthe purpose of analyzing dispari-ties in the distribution of the ben-efits of regional development andthe concentration of resourcebase and economic activities, therural and urban strata were fur-ther subdivided into smaller sub-strata.

2. The level of urbanization in thestudy area was analyzed.

3. The process of urbanization inthe study area was analyzed,based on the selected variables ofpush and pull factors by using amultiple regression model.

4. The location quotient (LQ) andGini coefficient (GC), indicatorsof spatial concentration and spa-tial disparities employed in thisstudy, were calculated forvariables related to regional agri-cultural development, regionaleconomy and regional develop-ment in public facilities.

5. Lorenz curves were constructedto represent the results of theanalysis.

6. The impacts of urbanization onagricultural development wereassessed.

7. Relevant hypotheses were tested.

Mathematical Models The mathematical models present-

ed include models to determine therate of change in population densityover time and over space, a model toanalyze the urbanization process,models for calculating location quo-tients (LQ) and Gini coefficients(GC), and models to compare meansof spatial and temporal characteris-tics.

Rate of Change in PopulationDensity Over TimeThe rate of change is commonlydefined as a change in any character-istic or substance over a period oftime. If the rate of change is definedas dR, the change in the characteris-tic or substance is defined as dX andthe time needed for the change isdefined as dT, the mathematical rela-tionship among these three variablescan be presented as :

δR = δX .........................................(1)δ T

Using this model, the rate of change inpopulation density in the study area canbe determined by dR – the rate of changein population density, dX – thedifference of population density in agiven time period, and dT – the timeperiod. For example, if population densi-ty in 1982 is k people/km2 and popula-tion density in 1992 is m people/km2,the annual rate of change in populationdensity between 1982 and 1992 is equalto (m-k) people/km2 divided by 10 years.

Rate of Change in PopulationDensity Over Space

The rate of change in populationdensity over space is predicted todecline as the distance between thearea and the central business districtincreases (Blair, 1991; Dicken andLloyd, 1990; Rees, 1970). The rate ofchange in the population densityover space is commonly predictedby a general model of a spatial decayfunction (Rees, 1970, p.277; Blair,1991, p. 373), presented as :

Then, the partial regression coeffi-cients—the slopes of the relationshipbetween Y and X1-k when the othervariables are held constant—can beused to determine which factors orvariables in the equation significant-ly determine the urbanizationprocess in the NCR.

However, to show comparablerates of change, all variables must beexpressed in the same measurementscale. This can be obtained by trans-forming the A coefficient into the coefficient, using the standardizedvariable value such as :

ZXk = (Xk – X)/Sx .......................... (9)

Where : ZXk is standard Z-score ofthe variable (Xk) value

X is mean value, andSx is standard deviation

Equation 8, then, can be transformedas equation 10 below :

ZY0.1-(n+m) = 0.1-(n+m) +

01.-(n+m) Xk + E ........................... (10)

Because we are dealing with Z-scores, the mean of every variable iszero so that the value of 0.1-(n+m)is zero. Consequently, equation 10can be expressed as equation 11below :

ZY0.1-(n+m) = 01.-(n+m) Xk+ E ........................(11)

To properly employ a linear modelof multivariate analysis (unbiasedestimates), all model assumptionsshould be satisfied. These (Johnston,1991) are: (i) linearity — partial plotof the dependent and independentvariable must depict a linearrelationship; (ii) the conditional dis-tributions of the residuals arenormal; (iii) the means of the condi-tional distribution — for every valueof X, the mean of (Yi – Ypred(i)) mustbe zero; (iv) homoscedascity —equal variances in the conditionaldistributions; (v) the value of eachobservation on the independent

βK

N M

=

+

∑1

β

β

15

Dx = D0e-bx ................................. (2)Where : Dx is population densityat a distance-xD0 is population density at theurban coree is a constant, andX is the distance from the villageto the capital city of the kabupaten(km)b is rate of change in populationdensity

In Equation 2 above, there is oneparameter, b, that can be predictedusing existing data from the studyarea. Therefore, in this study,parameter estimation is conductedby transforming equation 2 into alinear model, presented as :

ln Dx = ln D0 – bX ± E .................... (3)

Where : Dx is population densityin each village in the kabupaten inthe NCR

D0 is population density at thecapital city of the kabupatenX is the distance from the villageto the capital city of the kabupatenb is the rate of change in popula-tion density, andE is the standard error

Assuming that the capital city ofeach kabupaten in the NCR is alsothe center of development in thekabupaten makes it possible todetermine the rate of change of pop-ulation density over space in eachkabupaten. Using least square esti-mators (Neter and Wasserman, 1974;Johnston, 1991; Barber, 1988; Runyonand Haber, 1980; Ott, 1988), thevalue of the rate of change in popu-lation density (b), the expected valueof population density in the capitalcity of the kabupaten (D0) and thestandard error (E) of the regressionmodel (equation 3) in each kabupat-en can be determined through equa-tions 4, 5 and 6 below :Estimation of b:

b = ...... (4)

Estimation of In D0:

In D0 = - b ......................... (5)

Estimation of E:

E = ..(6)

Where n = number of samplevillages (1, 2, ..., i)

Urbanization AnalysisUrbanization is affected by two

groups of factors: rural push factorsand urban pull factors. If urbaniza-tion, measured as a rate of change inpopulation density, is expressed as Y,in the general multiple regressionmodel, the process of urbanizationin the NCR can be expressed as :

Y = f (rural push factors, urban pull factors) ................................... (7)

Suppose that rural push factorsconsist of all identified rural factorsaffecting urbanization in the NCR,expressed in the mathematical formsuch as S (Xi), for i = 1, 2, ..., N andurban pull factors, expressed as S(Xj), for j = 1, 2, ..., M; then, equation7 can be expressed in the multipleregression model such as :

Y = A0.1-(n+m) + A01.-(n+m) Xk+ E .............................................. (8)

Where : Y is the urbanizationprocess, expressed in rate of changein population density (people/km2/year)

Xk represents rural and urban factorsaffecting the urbanization process inthe NCR; all these factors are statedin the form of their rates of change,where k = 1, 2, ..., N+M A0.1-m are constants (value of Ywhen Xk = 0)A01.-m are partial regressioncoefficients, andE is an error term

K

N M

=

+

∑1

ΣΣ( ln

( )( )

Xi Dxi

Xi X Xi X

-X)( -lnDx)

− −

ln Dx X

Σ(ln )(ln )Dxi b bXi Dxi b bXi

n

− − − −−2

K

N M

=

+

∑1β

β

β

16

variable is independent of all of thevalues (autocorrelation); (vi) lack ofmeasurement error; and (vii) theresiduals from each partialregression equation are uncorrelated(collinearity).

Location Quotient (LQ) andGini Coefficient (GC)

The location quotient (LQ) is anindex for comparing an area’s shareof a particular activity with thearea’s share of some basic or aggre-gate socioeconomic activity. The LQfor a given activity for area i is theratio of the percentage of the totalregional activity in area i to the per-centage of the total base in area i. IfAi is equal to the level of the activityin area i and Bi is the level of thebase, then LQi is presented below(Barber, 1988, p.87):

LQi = .............................. (12)

The Gini coefficient (GC) or anindex of dissimilarity (Barber, 1988,p. 90) is used in this study tomeasure disparities in the regionaldistribution of developmentbenefits. Besides GC, there are sever-al alternative methods that can beused to measure dissimilarity, suchas range, relative mean deviation,variance and standard deviation ofthe natural log. (Cowell, 1977).

The range measures the differencebetween the highest and the lowestregional sample characteristics. Thismeasure ignores the spatial distribu-tion of the characteristics betweenthe extreme values and is sensitiveto outliers. The value of range fallsbetween zero (equally distributed)and n (fully concentrated). The rela-tive mean deviation looks at the entiredistribution, not only extremevalues. This measure may producethe same results from different dis-tributions on the same side of themean. The variance is a better mea-sure than the relative meandeviation because it is sensitive todifferences from the mean for allobservations. Its major disadvantage

is that a distribution could have alarger relative variation than anoth-er and still have a lower variance ifthe variation around the mean char-acteristic level is smaller than thecomparative distribution. The coeffi-cient of variation is a morecomprehensive measure than theprevious measures and is sensitiveto differences from the mean like thevariance, and independent of themean characteristic. It has, however,two weaknesses: the squaring proce-dure in its calculation procedure isarbitrary and it weighs differencesequally. The standard deviation of thenatural logarithm is the most usefulmeasure of inequity, especially ifone is interested in attaching greaterweight to extreme differences, suchas the lower income characteristicper capita. This measure is anappropriate one to analyze extremepoverty. However, it uses anarbitrary squaring procedure in itscalculation and it is seldom reportedso that it creates difficulties in com-parative studies of social equity(Cowell, 1977).

Among them, the GC is the mostcommon method used to measuredissimilarity of inequity (Smith,1982) and has better characteristicsthan the others. It is more sensitiveto the characteristic of differentials,independent of proportionalchanges in the characteristics andpopulation, avoids the arbitrarysquaring procedure and provides adirect measure of the characteristicdifferences (Cowell, 1977). Therefore,this study uses the Gini coefficient tomeasure disparities in the regionaldistribution of developmentbenefits.

The GC is measured using thelocation quotients approach (Smith,1982). The coefficient is derived asfollows:1. The variables used in the GC are

selected.2. Reference variables are selected

for use in the computation oflocation quotients.

3. Location quotients are calculatedfor all subregions (see equation12).

4. The subregions are rankedaccording to their locationquotient value.

5. The Gini coefficient is computedaccording to:

Gx = (YiXi+1 – Yi+1Xi) ........ (13)

Where Gx is the Gini coefficient ofvariable X

Xi is the cumulative share of vari-able X corresponding to subregion-i

Yi is the cumulative share of thereference variable corresponding tosubregion-i

Two types of Gini coefficient ratioscan be used:

1. To determine spatial disparitiesbetween subregions. When theGC ratio is close to unity, the setof subregions are almost identicalin spatial disparity. To the extentthat this ratio is different from1.0, regional differences exist.

2. To determine changes in inequal-ity over time. The ratio of theGCs can be calculated betweentwo points in time. If there is nosignificant change in the dispari-ty between the two, the ratio willbe close to 1.0. The higher theratio, the higher the spatialdisparity over time.

Sample Mean ComparisonTo compare regional

characteristics between 1982 and1992, the study used the t-testmethod for paired samples to testthe proposed hypotheses. FollowingBarber (1988), Netter andWasserman (1974) and Runyon andHaber (1990), the procedure andmathematical presentation of analy-sis can be summarized as follows:Suppose that the pair of valuesexpressed as X82i and X92i, thematched-pairs, is di = X82i – X92i; fori = 1, 2, ..., n. The hypothesisconcerning the value of samplemean difference,

Ai Ai

Bi Bi

/

/

ΣΣ

I

M

=

−

∑1

1

d

Sd n

−0

/

is tested using the t-test of

t = with degree of freedom

of n-1 for Sd =

Hypothesis testing is constructedas follow :

Hypothesis :H0 : µ1 – µ2 = µdHA : µ1 – µ2 is not equal to µd (twotailed), orHA : µ1 – µ2 < md (lower tail),implies H0 : µ1 – µ2 > µd, orHA : µ1 – µ2 > µd (upper tail),implies H0 : µ1 – µ2 < µd, or

Decision Rule :Reject H0 if t < -ta/2 or if t > ta/2 fortwo-tailed test with CF : µd + ta/2Sd/nReject H0 if t < -ta for lower tail testwith CF (confidence interval) : µd +ta Sd/nReject H0 if t > ta for upper tail testwith CF (confidence interval) : µd +ta Sd/n

To compare the characteristics ofrural and urban areas, this studyused the t-test for two independentsamples because the two populationvariances are not known. If the vari-ances of the two populations are thesame, sample variance 1 and samplevariance 2 are combined usingpooled variance,

S = , with

degree of freedom (df) of n1 + n2 -1and t-statistic of

t =

In these equations, X1 refers tosample 1 with number of sample n1and X1 refers to sample 2 with num-ber of sample n2, and the mean of Xis calculated by

=

17

If the variances of the two popula-tions are not the same, the t statisticused to test the hypothesis is

t = .

The degree of freedom for this caseis

df = .

Urbanization And ItsCausal Factors

The level of urbanization isdescribed on the basis of the rural-urban classification described earlier,followed by the analysis of the rateof change in population density orthe level of urbanization. In theanalysis of the urbanization process,regional pull and push factors aredescribed in detail so that specificfactors affecting the urbanizationprocess can be identified.

To analyze the urbanizationprocess, the study treats the NCR asa single entity without disaggregat-ing the region into administrativeunits of kabupatens or into severalagro-ecological zones. The NCR canbe treated as a single entity becauseall four kabupatens in the region,according to Firman (1992) andSoemarwoto et al. (1991), have simi-lar agricultural developmentproblems affected by urbanization.In addition, at the Jakarta-West Javaregional level, the NCR can be cate-gorized as one agro-ecological zonebecause irrigated lands for rice culti-vation are ubiquitous in the region(Soemarwoto, 1992), the region con-sists of similar types of farm manage-ment (Erwidodo, 1990), the area iscomposed of alluvial lowlands(Ongkosongo, 1979), and there are nosignificant regional differences inrainfall and temperature (averageyearly rainfall is about 1880 mm andtemperature ranges from 27.1 to 29.70 C (Ongkosongo, 1979; Ilahude,1979). Nevertheless, in determiningthe spatial rate of change in

population density, this study usesthe kabupaten as a basis of analysisbecause the capital city of each kabu-paten in the NCR is the center ofdevelopment in the kabupaten.

Urbanization in the NCRSimilar to Indonesia’s Central

Bureau of Statistics (CBS), this studydefines urbanization at the villagelevel. However, the criteria used toclassify rural and urban villages aredifferent. Here the process of rural-urban classification is described, fol-lowed by the analysis of the level ofurbanization. In the latter analysis,rates of urbanization of urbanvillages and urban populations arepresented and compared. Finally,results are presented in the form ofboth temporal and spatial rates ofchange in population densities.

Rural-Urban ClassificationUrbanization in Indonesia is

defined at the desa or village levelby categorizing the villages intorural villages and urban villages(Hamer et al., 1986; Firman, 1992).The urban population is defined asthe population in the villages catego-rized as urban. This study uses thesame method but a different criteri-on to classify the rural and urban vil-lages of the study area.

Several criteria can be used todefine urbanization and classifyrural and urban areas. The first is thelevel of regional economic progress,which is commonly approached byclassifying the aggregate measuresof the state of economic progress ofthe region (Williams et al., 1983). Thesecond is land rent, such as appliedin Schmid’s (1968) and Thrall’s(1987) studies in land conversionfrom rural to urban uses. The third ispopulation density, also a very com-mon measure used to classify ruraland urban areas (Goldberg andChinloy, 1984; Firman, 1992;Sharbatoghie, 1991; Hamer et al.,1986; Sutton, 1989; Obudho andWaller, 1976). Finally, two or morecriteria may be combined (Williamset al., 1983; CBS, 1988).

d

Sd n

−0

/Σ( )( )di d di d

n

− −−1

p2 ( ) ( )

( )( )

n s n s

n n

1 1 1 1

1 1 2 112

22− + −

− −

X X

Sp n n

1 2

1 1 1 2

−+/ /

XXi

n∑

X X

S n S n

1 2

1 212

22

−

+/ /

( / /

( / ) /( ) ( / ) /( )

S n S n

S n n S n n12

22

12 2

22 2

1 2

1 1 1 2 2 1

+− + −

Area Pop. density Number of villages Percent of villages(people/km2) 1980 1990 1980 1990

R1 < 1,000 760 635 69.15 57.78

R2 1,001-2,000 246 55 22.38 5.00

R3 2,001-3,000 47 325 4.28 29.57

R4 3,001-4,000 18 19 1.64 1.73

R5 4,001-5,000 10 28 0.91 2.55

RURAL < 5,000 1081 1062 98.36 96.63

U1 5,001-6,000 6 9 0.55 0.82

U2 6,001-7,000 5 5 0.46 0.46

U3 7,001-8,000 1 4 0.09 0.36

U4 8,001-9,000 2 6 0.18 0.55

U5 9,001-10,000 1 5 0.09 0.46

U6 > 10,000 3 8 0.27 0.72

URBAN > 5,000 18 37 1.64 3.37

NCR 1099 1099 100.00 100.00

18

Indonesia’s Central Bureau ofStatistics (1988) uses the fourth crite-rion. It classifies a village as an urbanvillage if population density exceeds5,000 people/km2, agriculturalhouseholds comprise 25 percent ofthe population or less, and eight ormore urban facilities are present.22

Although the Central Bureau ofStatistics uses multiple criteria, criticsargue strongly about defining suchcriteria. According to Rietveld (1988),all the criteria used to define urbanvillages in Indonesia are veryarbitrarily determined. Anagricultural household, for example,is defined as one where agriculture isthe main job of the head ofhousehold, though in reality there ismore than one working householdmember. The presence of an urbanfacility in the village is determinedwithout considering its quality; and,more importantly, village characteris-tics do not contribute significantly todetermining the degree of urbaniza-tion, such as the existence of an ele-mentary school in the village. InJava, the elementary school, especial-ly the public elementary school, ispresent in each village because gov-ernment standards dictate that eachvillage should have at least one ele-mentary school. Firman (1992, p. 97)has criticized these CBS criteria bystating that “these criteria aredifficult to apply because of inconsis-tency in result between them.”

Differing from Central Bureau ofStatistics (1988) and Firman (1992),this study uses population density asa single criterion to classify rural andurban areas in the NCR. Reasons forchoosing this single indicator are: (i)the absence of comprehensive mea-sures of economic progress availableat the village level; (ii) the lack offormal records or comprehensivestudies about land rent in the NCRthat can be used as a basis for classi-fication; (iii) the fact that populationand total area data are welldocumented at the village levelthrough national census data at 10-

year intervals and, therefore, popu-lation density for each village can bederived; and (iv) incomplete recordsof 1980 census data beyond data onpopulation and total area that can beaccessed via the Central Bureau ofStatistics, Jakarta, and in the Officeof Statistics in the kabupatens and inthe province. As a result of this lastreason, this study cannot use CentralBureau of Statistics criteria becausethe complete census data of 1990cannot be compared with the incom-plete census data of 1980.

Level of UrbanizationThis study uses a population den-

sity of 1,500 people/km2 as the cutoffbetween rural and urban villages inthe NCR. It finds that the use ofCBS’s criterion of 5,000 people/km2results in a very small number ofurban villages in the NCR. From atotal of 1099 villages in the NCR,only 18 villages in 1980 and 37

villages in 1990 could be categorizedas urban villages.

This represents only 1.64 percentof all villages in 1980 and only 3.37percent of all villages in 1990. Table 1and Figure 1 compare the number ofvillages categorized as rural andurban areas based on a village popu-lation density of 5,000 people/km2in 1980 and 1990.

The data (Table 1, Figure 1) showthat the use of a population densityof 5,000 people/km2 as thethreshold to differentiate rural andurban areas contradicts the assertionthat the NCR is the most urbanizedand industrialized region in Java(Soemarwoto, 1992; Firman, 1992;Hamer et al., 1986; Hill, 1992b),because the analysis shows that onlya very small portion of the NCR canbe categorized as urban villages.Hamer et al. (1986, p. 14), for exam-ple, calculated that 21 percent of theurban population in West Java (1980)

22 Facilities considered as urban facilities include primary school or equivalent, secondary school or equivalent, high school or equivalent, theater,hospital, maternity center, clinic, hardened or paved road, telephone/post office, market with permanent building, shopping center, bank, factory, restaurant, public electricity and party supply renting service.

Table 1. Number of villages in the Northern Coastal Region (NCR) of West Java by population density interval in 1980 and 1990.

Sources of data: 1980 and 1990 population censuses issued by the Central Bureau of Statistics,Jakarta, and the Offices of Statistics of the kabupatens Bekasi, Karawang, Subang and Indramayu.

19

was concentrated in the NCR(Firman, 1992). Since the calculationof urban population in Indonesia isbased on urban villages, the highurban population must reflect thehigh number of urban villages. Theresult will be much lower if the othertwo CBS criteria—agriculturalhouseholds and urban facilities—areincluded in the analysis. Therefore,adopting the CBS’s criteria impliesthat urbanization in the NCR is verylow and much lower in other regionsin Java, and it does not reflect thereality of urbanization in the NCR.In addition, Figure 3 shows that avillage population density of 5,000people/km2 is not the best thresholdto differentiate between rural andurban villages in the NCR becauseno clear distinction between ruraland urban villages is present.

This study suggests a cutoff of1,500 people/km2 to differentiatebetween rural and urban villages.Using this level of population densi-ty better reflects the level of urban-ization and the urbanization processin the NCR. The results of the analy-sis of the level of urbanization in theNCR are presented in Tables 2 and 3and Figures 2 and 3.

The population density levels ofrural and urban villages in the NCRin 1980 and 1990 are summarized inTable 2 and Figure 4. Rural andurban areas are subdivided further

into four subregions (R1-4 and U1-4)based on the population density.Subregions R1-4 to U1-4 indicate thedegree of urbanization from themost rural area (R1) to the mosturbanized area (U4). The levels ofrural and urban populations in 1980and 1990 and the levels ofpopulation in each subregion of therural and urban areas aresummarized in Table 3 and Figure 3.

Compared with Figure 1, Figure 2shows that a village population den-sity of 1,500 people/km2 provides aclear differentiation between ruraland urban villages in the NCR. Thisclassification represents the bestresult among the alternatives thathave been considered in this study.

Table 2 shows that there were 151urban villages (13.74 percent) in 1980and 223 (20.29 percent) in 1990. Thismeans that the number of urban vil-lages increased by 47.67 percentbetween 1980 and 1990. Table 3 andFigure 3, on the other hand, showthat urban population in the NCRincreased from 1,179,362 people(25.69 percent) in 1980 to 2,439,996people (38.75 percent) in 1990. Theaverage urban population increasedbetween 1980 and 1990 by 106.89percent, or 10.69 percent per year.This result shows that the level ofurban population in the NCR at25.69 percent in 1980 is higher thanthat of the level of urban population

in West Java of 21 percent ascalculated by Hamer et al. (1986).This result is expected because theurbanization process in West Java ismore concentrated in the NCR(Firman, 1992).

There are differences betweenurban village growth and urbanpopulation growth in the NCRbetween 1980 and 1990. In this peri-od, urban population in the NCRincreased by 10.69 percent per yearwhile the number of urban villagesincreased by 4.77 percent per year.This implies that urban populationgrowth in the NCR is determinednot only by the increase in urban vil-lages, thus changing the ruralvillages into urban villages, but alsoby changing demographic factors,rural and urban development,industrialization and other factors.According to Hamer et al. (1986, pp.iii and 1), “this growth is theoutcome of millions of individualhousehold and business decisions.”

Table 2 also shows that the num-ber of villages with a populationdensity of no more than 375people/km2 decreased significantlyfrom 194 villages (17.65 percent) in1980 to 91 villages (8.28 percent) in1990, a decrease of 53.09 percent or5.309 percent per year. This decreasealso occurred in the subregion —thenumber of villages with apopulation density of 376 to 750

Figure 1. Distribution of the villages in the NCR based on the village population density (source: Table 1).

U6 U5 U4 U3 U2 U1 R5 R4 R3 R2 R1

0100200300400500600700800

Subregions (U: Urban; R: Rural)

Number of Villages village number 90 village number 80

people/km2 decreased from 368(33.48 percent) in 1980 to 317 villages(28.84 percent) in 1990. In these sub-regions, the number of villagesdecreased by 13.86 percent between1980 and 1990. On the other hand, inthe other subregions the numberincreased between 0.576 and 6.5 per-cent per year in the period 1980-1990. These figures show that thenumber of villages in the NCR with

an indication of an agglomerativeeffect of urban population (Bhadraand Brandao, 1993).

Table 3, on the other hand, showsthat between 1980 and 1990, the pro-portion of rural populationdecreased in all subregions of therural area while the proportion ofurban population increased,especially in the urban fringe and inthe most populous subregions of theurban areas. These figures alsoimply that an increase in urban pop-ulation occurs mostly in the areascharacterized by previously higherpopulation density.

Temporal and Spatial Ratesof Change in PopulationDensity

The rate of change in populationdensity can be used as an indicatorof the concentration of populationover time and space. In this study,the temporal rate of change in thepopulation density (TRCPD) is con-sidered an indicator of the degree ofurbanization over a period of timebecause urban areas and rural areasare differentiated by population den-sity. And spatial rate of change inpopulation density (SRCPD)indicates the spatial differences inthe degree of urbanization from thecapital city of the kabupaten to thecountryside.

a population density of more than750 people/km2 increased in theperiod 1980-1990. Among theseregions, the urban areas show a larg-er increase than rural areas. Amongthe urban areas, the most populoussubregion (U4) had the highestincrease — 65 percent over 10 years.This implies that the highest urban-ization rate occurred in the regionwith the highest population density,

20

Figure 2. Distribution of rural and urban villages in the North Coastal Region of West Java in 1980 and 1990, based on a village population density of 1,500 people/ km2 as the cutoff to differentiate rural and urban villages (source: Table 2).

Area Pop. density Number of villages Percent of villages(people/km2) 1980 1990 1980 1990

R1 ≤ 375 194 91 17.65 8.25

R2 376-750 368 317 33.48 28.84

R3 751-1,125 247 321 22.47 29.21

R4 1,126-1,500 139 147 12.65 13.38

RURAL ≤ 1,500 948 876 86.25 79.71

U1 1,501-1,875 48 69 4.37 6.28

U2 1,876-2,250 27 34 2.46 3.09

U3 2,251-2,626 16 21 1.46 1.91

U4 ≥ 2,626 60 99 5.46 9.01

URBAN > 1,500 151 223 13.75 20.29

NCR 1,099 1,099 100.00 100.00

Table 2. Number of rural and urban villages in the NCR in 1980 and 1990 (classification based on a village population density of 1,500 people/km2 as the cutoff to differentiate rural and urban villages).

Sources of data: 1980 and 1990 population censuses issued by the Central Bureau of Statistics,Jakarta, and the Offices of Statistics of the kabupatens Bekasi, Karawang, Subang and Indramayu.

21

Temporal Rate of Change inPopulation Density

In this study, temporal rate ofchange in population density(TRCPD) is defined as the change inthe NCR’s population density overthe period 1980-1990 and isexpressed as people/km2/year. It iscalculated using equation 1, above.The results of the analysis arepresented in Tables 4 and 5, andFigure 4. Table 4 presents the meansof rate of change in population

density between 1980 and 1990(TRCPD) in rural and urban areas aswell as in each subregions in theNCR. These means are plotted inFigure 4 so that the patterns ofTRCPD from the most urbanizedregion (U4) to the most rural region(R1) can be easily seen. Table 5 pre-sents the result of the analysis ofvariance (ANOVA) of the means ofthe TRCPD in the NCR. It also pre-sents the result of the analysis ofmean comparison of the TRCPD in

each subregion in the NCR using theleast significant difference (LSD)tests of the multiple range tests.

Both Table 4 and Figure 4 showthat the TRCPD in the NCR is posi-tive between 1980 and 1990. Thisindicates that population densityincreased throughout the region.However, the increase in populationdensity is not evenly distributedbetween rural and urban areas andamong the subregions in rural andurban areas. The TRCPD decreasesfrom the most urbanized area (U4) tothe most rural area (R1). The meanof the TRCPD of the urban area is115.57 people/ km2/year and that inthe rural area is 12.76 people/ km2

/year. Statistically, the mean ofTRCPD of the urban area differsvery significantly from that of therural area, with a confidence level of99 percent (a = 0.01). The yearlyincrease in population density in the

urban area is 9.01 timesthat in rural area.Therefore, it can beinterpreted that urbaniza-tion in the NCR during theperiod 1980-1990 increasedmostly in areas with higherpopulation densities. Thissubstantiates the assertionby Henderson (1977) thatthe accumulation or move-ment of population in highdensity (core) regionsoccurs through a self-reinforcing process.

The accelerated growthin population densityover time is further sub-stantiated by comparing

the means of the TRCPD among sub-regions in the NCR (Table 5). Theanalysis shows that the TRCPD isthe highest in the most urbanizedarea (U4), at 221.22people/km2/year, and lowest in themost rural area (R1), at 5.10

Area Pop. density Number of villages Percent of villages(people/km2) 1980 1990 1980 1990

R1 ≤ 375 443,004 300,792 9.65 4.78

R2 376-750 1,234,220 1,234,220 26.89 20.22

R3 751-1,125 1,058,788 1,487,116 23.07 23.62

R4 1,126-1,500 674,900 795,385 14.70 12.63

RURAL ≤ 1,500 3, 410,912 3,856,372 74.31 61.25

U1 1,501-1,875 285,916 410,672 6.23 6.52

U2 1,876-2,250 162,206 203,129 3.53 3.23

U3 2,251-2,626 144,922 175,681 3.16 2.79

U4 ≥ 2,626 586,318 1,650,514 12.77 26.21

URBAN > 1,500 1,179,362 2,439,996 25.69 38.75

NCR 4,590,274 6,296,368 100.00 100.00

Table 3. Changing composition of rural and urban population in the NCR in 1980 and 1990.

Sources of data: 1980 and 1990 population censuses issued by the Central Bureau of Statistics,Jakarta, and the Offices of Statistics of the kabupatens Bekasi, Karawang, Subang and Indramayu.

Figure 3. Total population in each subregion in the NCR for 1980 and 1990 (source: Table 3).

0.01). The subregion U3 with aTRCPD of 61.10 people/km2/yearhas a significantly higher TRCPDthan U1 and R1-4 (a = 0.05) and alsohigher than U2 (a = 0.10). The subre-gion U2 with a TRCPD of 36.58 peo-ple/km2/year has a higher TRCPDthan R1-3 (a = 0.10). Finally,subregion U1, with a TRCPD of19.49 people/km2/year, has a higherTRCPD than R1-2 (a = 0.10).

The existence of the self-reinforc-ing process of population densitygrowth implies that—assuming thatspatial organization in the NCR isgoverned by spatial competition(Mabogunje, 1981; Mehretu, 1989)—people in the region prefer living inurban areas with high populationdensities than in rural areas withlow population densities. This mayimply that there exists a preferenceto locate in urban areas because of aperception of a higher quality of life.

This assertion will be vis-ited again in the analysisof the urbanizationprocess in the NCR.

Spatial Rate ofChange inPopulation Density

The spatial rate ofchange in population den-sity (SRCPD) is defined asthe spatial differences inthe degree of populationconcentration overdistance from the centerof the region (Gould,1972; Blair, 1991; Dickenand Lloyd, 1990). In thisstudy, SRCPD indicates

the spatial differences in the degreeof urbanization from the capital cityof the kabupaten to the countryside.SRCPD is calculated by using a gen-eral decay function model, a simplenegative exponential functiondescribed above.

This study assumes that themodel—which is successfullyapplied in the developed countries(Haggett, Cliff and Frey, 1977;Mehretu, 1989) to predict the decayof population density from the city

22

people/km2/year (see also Table 4).Further analysis using one-way

analysis of variance (ANOVA) andleast significant difference (LSD)multiple range tests show that verysignificant differences exist in themeans of the TRCPD among subre-gions in the NCR (Table 5).Comparing rural and urban areas, itseems that the TRCPDs in ruralareas are relatively consistently lowcompared with those in urban areas.

In rural areas, only R4 with aTRCPD of 20.02 people/km2/yearhas a higher TRCPD than that of R1and R2, which have TRCPDs of 5.10and 9.82 people/km2/year,respectively (confidence level of 90percent or a = 0.10). On the otherhand, in urban areas, U4 with aTRCPD of 221.22 people/km2/yearhas a very significantly higherTRCPD than U1-3 and R1-4(confidence level of 99 percent or a =

Area Population density Annual rate of change in population(people/ km2) density (people/km2/year)

R1 ≤ 375 5.1043

R2 376-750 9.8172

R3 751-1125 14.5185

R4 1,126-1,500 20.0236

RURAL ≤ 1,500 12.7631

U1 1,501-1,875 19.4939

U2 1,876-2,250 36.5776

U3 2,251-2626 61.0957

U4 ≥ 2,626 221.2178

URBAN > 1,500 115.5708

NCR 33.6240

Table 4. Mean of rate of change in population density between 1980 and 1990 (TRCPD) in the North Coastal Region (NCR) of West Java.

Sources of data: 1980 and 1990 population censuses issued by the Central Bureau of Statistics,Jakarta, and the Offices of Statistics of the kabupatens Bekasi, Karawang, Subang and Indramayu.

Figure 4. Mean of rate of change in population density between 1980 and 1990 in each of the NCR subregions (source: Table 4).

23

center to the countryside—alsoapplies in the NCR. Because no onecity in the NCR can be referred to asthe center of the whole NCR23, thisstudy uses the capital cities of thefour kabupatens as the city center.Consequently, the analysis of theSRCPD is conducted in eachkabupaten in the NCR; the regionalanalysis of the SRCPD in the NCR isconducted by assigning an averagevillage distance to the kabupatencapital city.

This research addresses the valueof b in each kabupaten and in the

whole region. In predicting suchparameters, the equation istransformed into a linear model.Using the least square method or theleast square estimators (Neter andWasserman, 1974; Johnston, 1991;Barber, 1988; Runyon and Haber,1980; Ott, 1988), the value of b, ln D0and E can be determined by usingequations 4, 5 and 6. The results ofthe analysis are presented in Table 6and Figure 3. Table 6 presents thevalue of b, ln D0 and E along withthe statistical measures related to thelinear model, such as the strength of

linear association (r, linearcorrelation), measure of goodness offit of the regression model (r2, coeffi-cient of determination) and signifi-cant testing of the strength of thevariable relationships (Snedector’sF-ratio). Figure 3 presents the resultof the linear relationship betweenthe natural log of a village’s popula-tion density and its distance fromthe capital city of the kabupaten. Thefigure plots only the positive valueof the natural log of the village’spopulation density by assigning

Analysis of Variance (ANOVA)

Source D.F. Sum of squares Mean squares F-ratio F-prob.

Among subregions 7 3911916.67 558845.24 43.33 0.00Within region 1091 14071540.23 12897.84Total 1098 17993456.90

Multiple range tests: LSD tests with a (a = 0.01; b = 0.05; c = 0.10)

Subregions R1 R2 R3 R4 U1 U2 U3 U

R1 -

R2 -

R3 -

R4 c c -

U1 c c -

U2 c c c -

U3 b b b b b c -

U4 a a a a a a a -

Table 5. Mean comparison of the temporal rate of change in population density (TRCPD) among subregions in the NCR using one-way ANOVA and least-significant difference (LSD) multiple range tests.

Sources of data: 1980 and 1990 population censuses issued by the Central Bureau of Statistics, Jakarta, and the Offices of Statistics of the kabupatensBekasi, Karawang, Subang and Indramayu.

23 There is no big city in the NCR, although this region is located between two of the biggest cities in Indonesia, Jakarta and Cirebon. However, there areat least four smaller cities in the NCR, which also happen to be the capital cities of the four kabupatens in the NCR.

Linear model: ln Dx = ln D0 – bX ± E

NCR Bekasi Karawang Subang Indramayu

b 0.044077 0.044255 0.048547 0.043955 0.041708

ln D0 1.867912 1.815106 1.807865 1.904647 1.888894

r 0.872140 0.846900 0.892590 0.915010 0.861320

r2 0.760620 0.717240 0.796720 0.837240 0.741870

Adjusted r2 0.760400 0.716040 0.796040 0.836580 0.741030

E 0.395100 0.482440 0.322270 0.286890 0.444110

F-statistic 3485.715 596.0956 1175.795 1275.720 885.1787

Sign. level24 0.0000 0.0000 0.0000 0.0000 0.000

population density. However, the values of the expect-

ed result—b or the SRCPD—are verylow in all regions, with values of0.0441, 0.0443, 0.0485, 0.0440 and0.0417 for the NCR, Bekasi,Karawang, Subang and Indramayu,respectively. This means that inBekasi, for example, for each kilome-ter farther from the capital city of thekabupaten, the population densitydeclines by 0.0443 people per squarekilometer. This implies that no signif-

icant difference existsin population densi-ty over distance fromthe capital city of thekabupaten.Therefore, the ranksize rule may not bethe probable distrib-ution of populationin the NCR. In otherwords, as Mehretu(1983) found in hisstudy of cities of sub-Saharan Africa, thesimple negativeexponential function,which has been avery successfulmodel for the devel-oped countries, is not

between a village’s population densi-ty and the distance from the capitalcity of the kabupaten in all regions isvery significantly strong, with a con-fidence level of more than 99 percent(see footnote 3). The values of thecoefficients of determination (r2) arealso very high in all regions —between 0.72 and 0.84. These indicatethat distance from the capital city ofthe kabupaten explains 72 to 84 per-cent of the variance of the villages’

numbers to the regression equations.All statistical measures used in this

analysis show that the model fits thesituation in the region, including allkabupatens. The values of linear cor-relation ® in all regions are very high,between 0.85 and 0.92, indicating avery strong predictive relationshipbetween a village’s population densi-ty and its distance from the capitalcity of the kabupaten. Furthermore,F-tests show that the relationships

24

Table 6. Rate of change in population density (b) over distance (SRCPD) from the capital city of the kabupaten predicted using a linear model with least square method.

Sources of data: 1980 and 1990 population censuses issued by the Central Bureau of Statistics, Jakarta, and the Offices of Statistics of the kabupatensBekasi, Karawang, Subang and Indramayu.

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

2.00

0 5 1 1 20 25 30 35 40

Distance from Capital City of the Kabupaten

NCR

Bekasi

Karawang

Subang

Indramayu

Ln

(pop

ulat

ion

dens

ity, p

eopl

e/km

2 )

Figure 5. Distance decay function of the population density from the capital city of kabupatens Bekasi, Karawang, Subang and Indramayu, and in the NCR in 1990 (source: Table 6).

24 F-test with significant level of a = 0.0000 means that the regression model fits the relationship with a confidence level of more than 99 percent.

25

tural land conversion into non-agri-cultural uses that affect the viabilityof agriculture in rural areas. Thisresearch indicates that in the past 10years, the proportion of urbanvillages in the NCR has increasedfrom 13.74 percent to 20.29 percent.Similarly, the urban populationincreased from 1,179,362 people to2,439,996 people — from 25.69 per-cent to 38.75 percent, an annualurbanization rate of 10.69 percentper year. It can be expected that sucha high rate of urbanization will neg-atively affect regional agriculturaldevelopment.

Operational DefinitionsThis study defines agricultural

development as any systematiceffort to (1) maintain and improvethe agricultural infrastructure; (2)promote farm landownershipthrough, among other things, reduc-tion of absentee landownership; (3)maintain and improve food securityof rural populations; (4) improve thewelfare of rural households, and (5)reduce rural-urban disparities, allwithout reducing the long-term pro-ductive capacity of the land resourcebase. The implicit assumption in thisdefinition is that rural food security,overall quality of life and socialequity can be enhanced by maintain-ing and improving the agriculturalresource base, strengthening itsownership structure and reducingthe dependency on absenteelandownership through comprehen-sive regional developmentstrategies.

This study analyzed changes inthe structure of landownership,changes in the degree of absenteelandownership, changes in agricul-tural prime land (or agriculturalinfrastructure), agricultural landconversion with its effects onagricultural household income,changes in the structure of ruralemployment and rural-urban dispar-ities in the period 1982-1992.

Other prominent impacts ofurbanization on agricultural devel-opment, such as the changing quali-ty of the agricultural resource base

tain the region as the nation’spredominant food producer and ahierarchical regional developmentstrategy. Urbanization and agricul-tural development affect each otherby direct or indirect linkages, includ-ing the underlying causal factors ofurbanization itself. Increased region-al urbanization on Java affects agri-culture by putting additionalpressure on the limited islandresource base, including increasedfood demand and the need toincrease productivity with a declin-ing resource base. At the same time,increasing agricultural activity stim-ulates urbanization by inducingregional industrial developmentthrough redistribution of laboramong rural and urban sectors. Inaddition, current government poli-cies undermine the viability of theagricultural sector and cause rural-urban disparities, forcing rural peo-ple to seek better economic opportu-nities in urban areas.

Urbanization is largely influencedby push factors from ruralagricultural areas, though pullfactors of urban areas anddemographic factors play a role aswell. For example, regionalurbanization is associated very sig-nificantly with an increase in agricul-

representative for the NCR.Beyond the statistical reliability of

the model, the actual value of thenatural log of a village’s populationdensity is plotted against thepredicted value derived from theanalysis (Figure 6). This plot showswhether the model provides an over-or underestimation of village popu-lation density.

The 45°-line in Figure 6 indicates aperfect model of prediction. If themodel is a good predictor, the actualvalues of the natural log of thevillage’s population density will bescattered close to this line. In thiscase, however, the actual values ofthe natural log of the village’s popu-lation density are concentratedbelow the line. This indicates thatthe model overestimates actual val-ues and does not represent theregion’s condition.

Urbanization andAgriculturalDevelopment

The relationship between urban-ization and agriculturaldevelopment in Indonesia is affectedby both national and regional devel-opment policies — a policy to main-

-4

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

10

-4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10

Actual value [Ln(population density), ppl/km2]

Pre

dic

ted

val

ue

[LN

(po

pu

lati

on

den

sity

), p

pl/k

m2]

Figure 6. The relationship between the actual values of the natural log of a village’s population density and its predicted values derived from the analysis.

Variable Percent of households with land Mean of paired Significantdifferences differences

1982 1992

Dryland :

0 HA 47.6639 56.6427 -8.9789 *****

< 0.5 HA 32.3098 30.8595 1.4502

0.5 – 1 HA 12.1532 8.5748 3.5784 *****

>1 – 2 HA 5.2080 3.1732 2.0348 *****

> 2 HA 2.1645 0.7380 1.4266 *****

Paddyland :

0 HA 43.4802 53.4057 -9.9255 *****

< 0.5 HA 27.4689 27.0698 0.3991

0.5 – 1 HA 16.4220 12.0536 4.3684 *****

>1 – 2 HA 9.2136 5.7011 3.5125 *****

> 2 HA 3.5350 1.6350 1.9000 *****

26

as a result of intensive land use pat-terns, soil and water pollutioninduced by urban-based agriculturalmanagement, and environmentalimpacts of fertilizers and pesticides –are not pursued in this study

Changes in the Distributionof Regional Landownership

The structural changes inlandownership of dryland and pad-dyland, both in the rural area in theNCR and in the NCR as a whole,were assessed at the village level bycomparing 1992 data with 1982 data.Statistical mean comparisons wereconducted using t-tests for pairedsamples. The results are summarizedin Table 7 and 8. Table 7 summarizesthe comparison between the percent-age of households with various landsizes (drylands and paddylands)within the NCR. Table 8 summarizesthe comparison between the percent-ages of households with variousland sizes (dryland and paddylands)in the rural areas within the regionin 1982 and 1992.

a. Changes in Landownership Distribution in the NCRThe structure of agricultural

landownership in the NCR between1982 and 1992 changed very signifi-cantly (Table 7). These changes putthe objectives of sustainable agricul-tural development at risk, especiallythose designed to improve the wel-fare of people engaged inagriculture, as a result of increaseddiseconomies of scale due to landparcelization. Analysis shows thatthe percentage of landlesshouseholds increased. Specifically,the percentage of households with-out agricultural dryland and paddy-land grew very significantly.Households without agriculturaldryland increased from 47.66 to56.64 percent in 1992, andhouseholds without paddylandincreased from 43.48 percent in 1982to 53.41 percent. Therefore, the num-ber of households withoutpaddyland grew faster than thenumber without dryland, increasingby 22.84 percent and 18.84 percent,respectively.

The increase in regional landlesshouseholds is not significantly asso-ciated with the decrease in the per-centage of households with < 0.5 hadryland or < 0.5 ha paddyland.Although households with < 0.5 hadryland and households with < 0.5ha decreased from 32.31 to 30.86 per-cent and from 27.47 percent in 1982to 27.07, respectively, this decrease isnot significant. This indicates thatthe percentage of subsistence farm-ers in the region has remained large-ly unchanged. Regional agriculturaldevelopment has, therefore, beenunable to reduce the percentage ofsubsistence farmers.

However, the increase in regionallandless households is very signifi-cantly associated with decreases inthe percentage of households with ≥0.5 ha dryland and ≥ 0.5 ha paddy-land during the period of analysis.Households with 0.5 to 1.0 hadryland and those with 0.5 to 1.0 hapaddyland decreased verysignificantly, from 12.15 percent to8.57 percent and from 16.42 percentto 12.05 percent, respectively.

Table 7. The differences in the distribution of household landownership in the NCR between 1982 and 1992.

Source: primary data (village survey)Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

27

Variable Percent of households with land Mean of paired Significantdifferences differences

1982 1992

Dryland :

0 HA 49.9700 58.4400 -8.4700 *****

< 0.5 HA 28.5488 27.6554 0.8933

0.5 – 1 HA 12.1075 8.5271 3.5804 *****

>1 – 2 HA 6.3692 4.3267 2.0425 ****

> 2 HA 2.5496 1.0708 1.4787 *****

Paddyland :

0 HA 47.4821 55.9400 -8.4879 *****

< 0.5 HA 26.9838 25.4217 1.5621

0.5 – 1 HA 13.6300 10.9513 2.6788 ****

>1 – 2 HA 9.4188 5.7129 3.7058 ***

> 2 HA 3.4096 1.7192 1.6904 *****

Table 8. Differences in the distribution of household landownership in the rural areas of the NCR between 1982 and 1992.

These decreases are the greatestamong other land divisions. For thesame 10-year period, householdswith 1.0 to 2.0 ha of dryland or pad-dyland decreased very significantly,from 5.21 percent to 3.17 percent andfrom 9.21 percent to 5.70 percent,respectively. The decrease in the per-centage of households with this hec-tarage of paddyland is greater thanthose with dryland (39.16 percentand 38.11 percent, respectively).Finally, the percentage of householdswith more than 2.0 ha of drylanddecreased less than paddyland.Households with more than 2.0 ha ofdryland decreased by 65.74 percent(from 2.16 percent in 1982 to 0.74percent in 1992). Similarly, thehouseholds with more than 2.0 hapaddyland decreased by 53.67 per-cent (from 3.54 percent to 1.64percent).

b. Changes of Landownership Distribution in Rural AreasChanges in the agricultural

landownership distribution in ruralareas of the NCR during the 1982-

1992 period are of the same magni-tude as those in the NCR as a whole.The number of landless householdsincreased and the households withagricultural landholdings decreased.No significant differences existbetween the percentage ofhouseholds with < 0.5 ha drylandsand those with < 0.5 ha ofpaddyland in the period 1982-1992.Results of the analysis of changes inthe distribution of agriculturallandownership in the rural areas aresummarized in Table 8.

Table 8 shows that the percentagesof households without agriculturaldrylands and paddylands increasedvery significantly in the period 1982-1992, by 16.95 percent and 17.82 per-cent, respectively. However, house-holds with < 0.5 ha agricultural dry-land or paddyland showed no statis-tically significant increase.Households with 0.5 to 1.0 ha agri-cultural dryland decreased very sig-nificantly, from 12.11 to 8.53 percent,a relative decrease of 29.56 percent.Households with 1.0 to 2.0 ha

agricultural dryland decreased sig-nificantly (32.03 percent). Also, thepercentage of households with > 2.0ha dryland decreased verysignificantly (58.04 percent).

Similarly, the percentage of house-holds with 0.5 to 1.0 ha paddylanddecreased very significantly (19.66percent). The percentage ofhouseholds with 1.0 to 2.0 ha paddy-land decreased significantly (39.38percent) and the percentage ofhouseholds with more than 2.0 hadryland decreased very significantly(49.56 percent).

The analysis shows a very signifi-cant increase in agricultural landparcelization during the period1982-1992, characterized by anincreasing percentage of landlesshouseholds. Agricultural landparcelization puts agricultural devel-opment at risk because it threatensthe welfare of the populationengaged in agriculture because ofdiseconomies of scale in agriculturalenterprises (Toner, 1979; Dunford,1981); and Lockeretz, 1986).

Source: primary data (village survey)Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

28

Moreover, the combined effects ofincreased absentee landownershipand increased agricultural land con-version in rural areas underminesthe role of the region as the rice bowlof Indonesia. This effect may be miti-gated through future consolidationof agricultural lands.

Absentee Landownership inRural Areas and the NCR

Regional urbanization is common-ly accompanied by increasing pricesof agricultural land. With priorknowledge of regional land useplans, land speculators can acquireagricultural land proposed for urbandevelopment. This increases the hec-tarage of land in rural areas ownedby people from outside rural areas.In the NCR, the increased hectarageof land owned by outsiders, howev-er, is determined not only byincreased land speculation, but alsoby other factors such as increasedinvestment in agricultural land as ahedge against inflation.

In sociological terms, the landowned by outsiders is usuallyreferred to as absentee land and iscontrolled by absentee landowners.

In the NCR, parallel terms such as“tanah gontai” and “tuan tanah” (or“spekulan” to indicate land specula-tors) are commonly used.

The study describes changes in thedegree of absentee landownership inthe rural areas and the region as awhole during the period 1982-1992,including changes in the hectarage ofabsentee lands and the number ofabsentee landlords. It does not, how-ever, differentiate between the causesof absentee lands, i.e., land specula-tion or other socioeconomic reasons.

a. Absentee Landownership in the NCRChanges in absentee landowner-

ship in the NCR between 1982 and1992 are summarized in Table 9,which presents absentee ownershipof dryland, paddyland and totalland.

The total absentee lands and thenumber of absentee landlords in theregion increased very significantly.Average absentee lands in theregion, assessed at the village level,increased by 193.20 percent, or 19.32percent annually. The average num-ber of absentee landlords increasedby 12.30, an increase from 28.43 in

1982 to 40.73 in 1992. Therefore, theaverage number of absenteelandlords increased by 4.33 percentannually. The total hectarage ofabsentee lands increased faster thanthe number of absentee landlords,indicating a concentration in absen-tee landownership.

Comparing the total hectarage ofabsentee lands of dryland and pad-dyland shows that increased absen-tee hectarage can be attributed to anincrease in absentee dryland.Average absentee dryland in theregion increased very significantly,from 12.45 ha in 1982 to 87.89 ha in1992. This means that the totalabsentee dryland in the NCRincreased by 60.59 percent annually!Average absentee paddyland, mean-while, increased very significantly,by 14.40 ha. This means that the totalabsentee paddyland in the NCRincreased by 42.29 percent duringthe period 1982-1992, or 4.23 percentannually.

The pattern of increase in absenteedryland is accompanied by a highernumber of absentee landlordsacquiring dryland than acquiringpaddyland. The average number ofabsentee landlords with dryland

Lands owned by and the number Mean of Significantof absentee landlords

Total in the villagepaired differences

in the village differences1982 1992

ha ha haTotal absentee land:1. Dryland 12.4530 87.8927 -75.4398 *****2. Paddyland 34.0455 48.4470 -14.4016 *****3. Dryland and paddyland 46.4985 136.3397 -89.8412 *****

no. no. no.Total absentee landlords with:1. Dryland 3.3409 11.0000 -7.6591 *****2. Paddyland 25.0909 29.7273 -4.6364 ***3. Dryland and paddyland 28.4318 40.7273 -12.2955 *****

Table 9. Regional differences in the hectarage of absentee lands and the total number of absentee landowners or absentee landlords between 1982 and 1992.

Source: primary data (village survey)Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

29

increased by 22.93 percent annually.Meanwhile, the average number ofabsentee landlords with paddylandincreased only 1.85 percent annually.

While it is more profitable toinvest in paddyland than in dryland,the high increase in absenteedryland in the NCR might be attrib-uted to increased regional land spec-ulation because rural land allocatedfor urban uses is rarely irrigatedland. This is especially importantgiven the fact that existing GOI poli-cies prohibit conversion ofpaddyland to non-agricultural uses.

b. Absentee Landownership in the Rural AreasChanges in absentee landowner-

ship in the rural areas of the NCRduring the period are more stagger-ing than those in the region as awhole. Absentee landownership inrural areas for the period 1982-1992is summarized in Table 10. It showsthat the degree of absentee landown-ership in the rural areas is muchgreater than that in the whole NCR.

Tables 10 and 11 show that in 1982the average hectarage of absenteelands and the number of absenteelandlords at the regional level were

areas can be attributed more to theincrease in absentee dryland, whichincreased by 63.81 percent annually.Absentee paddyland, meanwhile,increased by 2.98 percent annually.

Increases in absentee dryland areaccompanied by a higher number ofabsentee landlords acquiringdryland rather than paddyland.However, the average number ofabsentee landlords with paddylandincreased insignificantly, indicatingthat absentee paddyland in ruralareas is also becoming more concen-trated in absentee ownership.

Agricultural LandConversion in Rural Areasand the NCR

According to Schmid (1968) andFirman (1992), regional urbanizationis accompanied by increased demandfor land to support residential andindustrial development. Since vacanturban land is scarce and typicallymore expensive, agricultural land iscommonly converted to non-agricul-tural use. Therefore, it can be expect-ed that agricultural land in the NCRis subject to conversion.

46.50 ha and 28.43, respectively.These are smaller than in the ruralareas, where they were 67.88 ha and40.38, respectively. In 1992, thefigures for the NCR are 136.34 haand 40.73; in the rural areas they are198.86 ha and 54.42, respectively.

Changes in landownership inrural areas, however, are similar tothose in the region. Table 10 showsthat total absentee lands and thenumber of absentee landlords inrural areas increased verysignificantly between 1982 and 1992(confidence level of 99 percent).

Average absentee village landincreased by 130.98 ha. This meansthat total absentee lands in the ruralareas increased by 19.30 percentannually. The average number ofabsentee landlords in rural areas,meanwhile, increased by 14.04, or3.48 percent annually. As at theregional level, the increase in absen-tee land is larger than the increase inthe number of absentee landlords,indicating a concentration of absen-tee landownership.

Comparing the total hectarage ofabsentee lands (dryland and paddy-land), Table 10 shows that the totalincrease in absentee land in rural

Lands owned by and the number Mean of Significantof absentee landlords

Total in the villagepaired differences

in the village differences1982 1992

ha ha haTotal absentee land:1. Dryland 18.2138 134.4054 -116.1920 *****2. Paddyland 49.6667 64.4563 -14.7896 *****3. Dryland and paddyland 67.8805 198.8617 -130.9812 *****

no. no. no.Total absentee landlords with:1. Dryland 2.2083 12.3750 -10.1667 *****2. Paddyland 38.1667 42.0417 -3.8750 ***3. Dryland and paddyland 40.3750 54.4167 -14.0417 *****

Table 10. The differences in the absentee landownership and the number of absentee landowners (absentee landlords) in the rural areas of the NCR between 1982 and 1992.

Source: primary data (village survey)Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

30

land is weighted by the number ofvillage population.

b. Agricultural Land Conversion inRural Areas

Changes in total agricultural landbetween 1982 and 1992 in rural areasof the region are summarized in(Table 12).

Irrigated land and agriculturaldryland conversions aresummarized in total hectarage andweighted by the village population.In rural areas, the village’s averageirrigated land decreased very signifi-cantly, while the average agriculturaldryland decreased insignificantly.This indicates that agricultural landconversion affects mostly irrigatedland. This represents an averagedecrease of 2.50 percent annually.This decrease is also very significantwhen the total hectarage of irrigatedland is weighted by the number ofvillage residents. The average agri-cultural dryland conversion is, onthe other hand, insignificant.Weighted by number of village resi-dents, however, this decrease is sig-nificant.

It is clear that land conversionanalysis in rural areas and in the

region as a whole shows that irrigat-ed land has been sacrificed in theprocess of urbanization. This processundermines the future sustainabilityof regional agriculture, especially inmaintaining its significant contribu-tion to the national policy of foodself-sufficiency.

Impacts of Agricultural LandConversion on HouseholdIncome

At the regional level, the increasein irrigated land conversion is risk-ing the NCR’s capacity to maintainits key role as food supplier. At thehousehold level, this does not neces-sarily mean that those directly affect-ed by irrigated land conversion willsuffer diminished welfare. Therefore,this study assessed the impacts ofirrigated land conversion on the wel-fare of the households involved inirrigated land conversion. Usingdata from the Project of LandConversion in Java, Center for Agro-Economic Research (PAE), this studycompared the welfare of farminghouseholds before and after irrigat-ed land conversion. The data repre-sent on-farm, off-farm and total

This study assesses the changes intotal agricultural lands in the ruralareas and in the region as a wholebetween 1982 and 1992 to determinewhether irrigated land is mostly con-verted to non-agricultural uses. Thisquestion addresses the policy objec-tive to preserve the region’s impor-tance as the “rice bowl” ofIndonesia.

a. Agricultural Land Conversion in the NCRResults of the analysis of changes

in the total agricultural landbetween 1982 and 1992 in the NCRare summarized in Table 11. It showsirrigated land and agricultural dry-land conversions, both in total hec-tarage and weighted hectarage bynumber of the village population.Results indicate that average irrigat-ed land and the average agriculturaldryland decreased significantly. Italso indicates that agricultural landconversion affects mostly irrigatedland, the pillar of agriculturalproductivity of the region. Averageirrigated land hectarage decreasedby 2.56 percent annually. Thisdecrease is also very significantwhen the total hectarage of irrigated

Types of agricultural land Mean of Significantconverted into

Total in the villagepaired differences

non-agricultural uses differences1982 1992

ha haIrrigated land:1. Dryland 285.6945 212.5166 73.1778 *****2. Total hectarage for each 58.2595 34.1545 24.1050 *****

1,000 village population

ha haAgricultural dryland:1. Total hectarage 67.3977 49.1259 18.2718 **2. Total hectarage for each 14.7352 8.0759 6.6593 ***

1,000 village population

Table 11. Differences in the total land (ha) and total land for each 1,000 in village population (ha/1,000 village population) for both irrigated land and agricultural dryland in the NCR between 1982 and 1992.

Source: primary data (village survey)Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

31

income of 52 households involved inirrigated land conversion in the NCRover the past 10 years. Householdwelfare is measured by total house-hold income per year and convertedinto the 1992 present values using anaverage inflation rate of 9 percent.With this standardized value ofincome, the level of income beforeand after land conversion can becompared using the t-tests methodfor paired samples. Results of theanalysis are summarized in Table 13.

It includes average householdincome on-farm and off-farm, andtotal income before and after landconversion.

Table 13 shows that totalhousehold income (adjusted forinflation) after land conversion isvery significantly higher than thatbefore land conversion. Incomeincreased from Rp. 3,037,196.58 toRp. 4,178,545.29, an increase of 37.58percent. This indicates that the eco-nomic welfare of the households

increased almost 4 percent per yearafter conversion of irrigated land.

Comparing total income with on-farm and off-farm income, it showsthat both on-farm and off-farmhousehold incomes are significantlyhigher after land conversion. Thefact that on-farm income is higherafter land conversion is surprisingunless households involved acquireadditional agricultural land afterland conversion. Further analysisconducted in this study verifies this

Types of agricultural land Mean of Significantconverted into

Total in the villagepaired differences

non-agricultural uses differences1982 1992

ha haIrrigated land:1. Dryland 382.3441 286.5920 95.7521 *****2. Total hectarage for each 80.9563 47.1675 33.7888 *****

1,000 village population

ha haAgricultural dryland:1. Total hectarage 78.3896 55.9579 22.43172. Total hectarage for each 19.7113 9.9446 9.7667 **

1,000 village population

Table 12. Differences in the total land (ha) and total land for each 1,000 in village population (ha/1,000 village population) for both irrigated land and agricultural dryland in the rural areas of the NCR between 1982 and 1992.

Source: primary data (village survey)Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

Sources of Mean of household income Mean of paired Significantincome (Rp./year) differences differences

Before land conversion After land conversion

On-farm 1,576,587.62 2,102,767.88 -526,180.26 ****

Off-farm 1,460,608.96 2,075,777.40 -615,168.44 *****

Total 3,037,196.58 4,178,545.29 -1,141,348.71 *****

Table 13. Differences in on-farm, off-farm and total incomes (Rupiah, Rp.) of farming households before and after irrigated land conversion during the period 1982-1992, adjusted for inflation .

Source: Household survey conducted by the Project of Agricultural Land Conversion in Java, the Center for Agro-Economic Research (PAE), 1993. Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

32

suspicion. In fact, average householdirrigated land significantly increasedfrom 1.0665 ha per household beforeland conversion to 1.2628 ha perhousehold after land conversion, anincrease of 0.1962 ha per household.This indicates that households usedthe returns from land conversion toinvest in other irrigated lands.

An additional question pursued inthis study is whether thehousehold’s decision to convert landis economically rational. If so, it maybe expected that regional land con-version is primarily dictated by mar-ket prices and that government poli-cy does not significantly affecthousehold decisions regarding landconversion.

The decision is economically ratio-nal if on-farm incomes ofhouseholds not involved in landconversion are insignificantly differ-ent from the after land conversionon-farm incomes of those involvedin land conversion. For the purposeof this analysis, on-farm income afterland conversion of those householdsinvolved in land conversion is com-pared with on-farm income ofhouseholds not involved in landconversion using the t-test for twoindependent samples. The result ofthe analysis is presented in Table 14.

Table 14 shows that no significantdifference exists in mean on-farmhousehold income before and afterland conversion. This indicates thatfarm households in the NCR willconvert their land if it is economical-ly profitable; otherwise, they willnot. This implies also that it is

difficult to restrict agricultural landconversion unless the governmentpromotes indirect policy measures toenhance farm household incomesand the profitability of agriculturalenterprises.

Structural Changes inRegional and Rural VillageEmployment

Modernization theory asserts thaturbanization is accompanied byregional structural employmentchanges. That is, regional urbaniza-tion is associated with a decrease inthe percentage of the populationemployed in the agricultural sectorcompared with industrial andservice sectors.

To test this assertion, structuralchanges of regional and ruralemployment in the NCR areassessed at the village level duringthe period 1982-1992. The primaryfocus of analysis is to determine ifthe percentage of people employedin the agricultural sector decreasedand whether an identical trend waspresent in rural areas. As previously,a paired t-test is used and results aresummarized below (Table 15).

Various regional and ruralstructural changes in employmentduring the period 1982-1992 are iden-tified. Because the analysis isconducted at the village level, thetypes of employment described inthis study are somewhat differentfrom those usually used to analyzeregional structural employment inthe regional economic literature.Village employment, in this study, is

classified on the basis of the largestshare (more than 50 percent) ofemployment contributing to totalindividual income. Eight types ofemployment are identified : 1. Agricultural labor — percentage

of the village population provid-ing labor input to farm enterpris-es.

2. Farmers — percentage of the vil-lage population engaged inowner-operated agriculturalenterprises.

3. Sakap/sharecroppers — percent-age of the village populationengaged in agriculture througheither land leases orsharecropping.

4. Manufacturing labor — percent-age of the village populationemployed in the manufacturingsector.

5. Small-scale enterprises —percentage of the village popula-tion engaged in subsistence,small-scale trade without laborinputs from outside thehousehold.

6. Government officials — percent-age of the village populationemployed in the public sector.

7. Private enterprises — percentageof the village populationemployed in the local businesssector using outside labor input.

8. Other — percentage of the villagepopulation engaged in otheremployment.

At the regional level, Table 15shows that village employment in1982 consisted of 22.49 percent

Mean of household income Mean of Significant(Rp./year) differences differences

After land conversion No land conversion

On-farm 2,102.767.88 1,564,273.75 538,494.13

Table 14. Differences between after land conversion on-farm income of the households involved in land conversion and on-farm income of those not involved in land conversion, 1992.

Source: Household survey conducted by the Project of Agricultural Land Conversion in Java, the Center for Agro-Economic Research (PAE), 1993. Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

33

agricultural labor, 24.03 percentfarmers, 20.16 percent sharecroppers,11.95 percent manufacturing labor,7.53 percent small-scale enterprises,4.45 percent government officials,4.02 percent private enterprises and5.38 percent other. This employmentstructure is very significantly differ-ent in 1992, at which time villageemployment was 22.27 percent agri-cultural labor, 12.23 percent farmers,14.78 percent sharecroppers, 22.42percent manufacturing labor, 11.09percent small-scale enterprises, 6.16percent government officials, 5.70percent private enterprises and 5.35percent other.

Although there is no significantchange in the percentage of agricul-tural labor, the percentages of farm-ers and sharecroppers decreasedvery significantly in the region

during 1982-1992 — 49.11 and 26.67percent, respectively .

Employment in manufacturing,small-scale enterprises, the govern-ment sector and the privateenterprise sector, meanwhile, grewvery significantly. Employment inmanufacturing increased by 87.62percent, small-scale enterpriseemployment by 47.28 percent andemployment in the government sec-tor by 38.43 percent. Finally, privateenterprise employment grew by41.79 percent.

These figures show that at theaggregate regional level of the NCR,decreased employment in theagricultural sector is accompaniedby a relatively high increase inemployment or an employment shiftto the manufacturing sector. It

means, indeed, that urbanization inthe NCR is accompanied by a struc-tural employment change, as assert-ed in modern urbanization theory.

On the other hand, the structuralchange in employment in the ruralareas of the NCR is of a differentmagnitude than in the region as awhole. In rural areas, villageemployment in 1982 was 28.59 per-cent agricultural labor, 31.15 percentfarmers, 26.10 percentsharecroppers, 2.53 percentmanufacturing labor, 7.14 percentsmall-scale enterprises, 1.75 percentgovernment sector, 0.76 percent pri-vate enterprises and 1.99 percentother. This structure is significantlydifferent from that in 1992, when vil-lage employment was 34.89 percentagricultural labor, 17.99 percentfarmers, 23.00 percent

Types of employment Percent of population Mean of Significantpaired differences

1982 1992 differences

In the NCR :

Agricultural labor 22.4864 22.2720 0.2143

Farmer (operator) 24.0291 12.2289 11.8002 *****

Sharecroppers (sakap) 20.1564 14.7836 5.3727 *****

Manufacturing labor 11.9532 22.4200 -10.4668 *****

Small-scale enterprise 7.5334 11.0880 -3.5545 *****

Government official 4.4484 6.1586 -1.7102 *****

Private enterprise 4.0159 5.7039 -1.6880 *****

Others 5.3780 5.3452 0.0327

In the rural areas of the NCR :

Agricultural labor 28.5929 34.8946 -6.3017

Farmer (operator) 31.1454 17.9908 13.1546 *****

Sharecroppers (sakap) 26.1000 22.9992 3.1008

Manufacturing labor 2.5258 6.2729 -3.7471 *****

Small-scale enterprise 7.1371 11.8246 -4.6875 ****

Government official 1.7463 3.1683 -1.4221 ****

Private enterprise 0.7646 1.2521 -0.4875 *****

Other 1.9883 1.5996 0.3887

Table 15. Differences in the regional and rural area’s employment structure between 1982 and 1992.

Source: primary data (village survey)Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

34

sharecroppers, 6.27 percentmanufacturing labor, 11.82 percentsmall scale enterprises, 3.17 percentgovernment sector, 1.25 percent pri-vate enterprises and 1.60 percentother.

Different from the regional aggre-gate, it shows a significant increasein the percentage of agriculturallabor in the rural areas (22.04percent). The percentage of farmersdecreased very significantly (by42.25 percent). No statistically signif-icant decrease is apparent in the per-centage of sharecroppers during thesame period.

Employment in manufacturingand private enterprises in rural areasgrew very significantly. Meanwhile,employment in small-scaleenterprises and the government sec-tor grew significantly, as well.Manufacturing employmentincreased by 147.73 percent, small-scale enterprise employment grewby 65.55 percent and governmentemployment increased by 81.14 per-cent. Finally, private enterpriseemployment increased from 0.76percent in 1982 to 1.25 percent in1992, a growth of 64.47 percent.

Five fundamental differencesbetween rural and regional structur-al changes of employment exist.First, rural employment is still domi-nated by agricultural employment.Second, the percentage of agricultur-al labor in rural areas increased sig-nificantly (22.04 percent). This ruralincrease is much higher than that inthe region (0.95 percent), a case ofconcentration of agriculturalemployment. Third, althoughrelative employment in the manufac-turing sector in rural areas increasedvery significantly (147.73 percentand 87.62 percent, respectively), itsabsolute value is much less than thatof the region. Fourth, the increase insmall-scale enterprises is higher inrural areas than in the region (65.55percent and 47.28 percent,respectively). Finally, theemployment ratio in privateenterprises in the rural areas is muchless than that in the region as awhole.

These figures imply that at theregional level a decrease in agricul-tural sector employment is accompa-nied by a significant increase inmanufacturing employment andsmall-scale enterprises, and adecrease in the percentage of self-employed farmers is not followed byan increase in employment in theagricultural sector, an indication thatmanufacturing industries and small-scale enterprises have the capacity toabsorb farmers displaced from theagricultural sector. Meanwhile, forrural areas the figures imply that adecrease in employment in the agri-cultural sector is not followed by asignificant increase in theemployment in the manufacturingsectors, and a decrease in thepercentage of self-employed farmersis followed by an increase inemployment as agricultural laborand small scale-enterprises, an indi-cation that displaced farmers in ruralareas are either being employed asagricultural laborers or entering thelabor force of small-scale enterprises.

Regional Developmentof the Northern CoastalRegion (NCR) of WestJava

Regional development of the NCRcannot be separated from regionaldevelopment of Jakarta-West Java asa whole, in which West Java hasbeen designated as the hinterland ofJakarta. In this regional developmentframework, sectoral development,especially industrial development —which is deemed saturated in Jakarta— is to be promoted in West Java(Hamer et al., 1986). This indicatesthat development in the regionfavors areas with betterinfrastructure and proximity to theseat of government. Because theNCR has a better infrastructure andproximity to Jakarta, the industrialdevelopment in West Java is concen-trated in this region (Hill, 1992b).

This study confirms Hill’sassertion that the government policyto subsidize medium- and large-

scale industries results in a higherdegree of industrialization in theregion. Instead of calculating thevalue added, this study assesses theincreased number of industries dur-ing the period 1982-1992. Theincreased number of industries indi-cates a higher degree ofindustrialization in the NCR.

Table 16 summarizes the result ofmean comparison in the number ofindustries between 1982 and 1992. Inthis analysis, the industry is catego-rized as: (i) large scale industry, i.e.,an industry with 100 or moreemployees; (ii) medium-scale indus-try, i.e., an industry with 20 to 99employees; and (iii) small-scaleindustry, i.e., an industry with 5 to19 employees. Changes in thedegrees of regional industrializationare assessed at the village level.

Table 16 shows that the villages’average number of large-scale indus-tries increased very significantlyfrom 0.09 in 1982 to 4.41 in 1992, anincrease of 4.3182 or 4750.50 percentduring the period, or 475.05 percentannually . The figure also increasedvery significantly after the numberof large-scale industries is weightedby the village population. Thevillages’ average number ofmedium-scale industries increasedsignificantly from 0.7273 in 1982 to2.6136 in 1992, an increase of 1.8864or 259.36 percent during the period,or 25.94 percent annually. The levelof significance of the increased num-ber of medium-scale industries is thesame after it is weighted by villagepopulation. This study also findsthat the small-scale industries in theNCR increased very significantlyduring this period, from 6.0909 in1982 to 11.0455 in 1992, an increaseof 4.9545 or 81.34 percent during theperiod, or 8.13 percent annually.

Along with the regional develop-ment of Jakarta-West Java, whichresulted in the increased number ofindustries of all scales, the NCR isalso affected by the regionaldevelopment of West Java.

The regional development masterplan of West Java was developed in1974 (Government of West Java,

35

1990) and is updated every fiveyears along with the five-year planfor national development. Accordingto this master plan, regionaldevelopment in West Java adopts afunctional integration rather than aterritorial integration as its strategyof regional planning with “growthcenters” as the principal instrumentof its spatial policy. In this study, agrowth center is defined as “inducedurbanization through a combinationof direct public investments and cap-ital subsidies to private enterprise”(Friedman and Weaver, 1979, p.6).Functional integration, meanwhile,is formed on the basis of mutual self-interests among regions. Giveninequalities at the start, a functionalintegration is always hierarchicalwith power accumulating at the top.On the other hand, territorialintegration is derived from commonbonds of social order forged by his-tory within a given place. Althoughthey will also be characterized byinequalities of power, territorial rela-tionships are tempered by mutualrights and obligations that the mem-bers of a territorial group claim fromeach other (Friedman and Weaver,1979; Friedman, 1988).

Such a regional developmentstrategy is biased toward the advan-tage of urban development(Friedman, 1988) and characterizedby large investment in the urban-

industrial sector (Rondinelli, 1985).Two important problems are associ-ated with urbanization driven bysuch a regional development strate-gy. First, agricultural developmentwill be negatively affected by urban-ization. Second, rural areas wouldtake part in the general process ofgrowth diffusion only to the extentthat they were subject to the impactof the urban economy (Friedmanand Weaver, 1979). In other words,the urbanization process is not asso-ciated with social urbanization andincreasing opportunities fordevelopment in rural areas(Mehretu, 1989). Therefore, it can beexpected that there are rural-urbandisparities in the distribution of thedevelopment benefits as a result of acumulative causation process(Myrdal, 1957) or self-reinforcingprocess.

Rural-Urban DisparitiesRural-urban disparities, in this

study, are defined as an uneven dis-tribution of goods, resources,incomes and services between ruraland urban areas of the NCR(Matthews, 1983). This study doesnot assume that rural-urban dispari-ties in the region are created by lackof resources in the rural areas andthe inability to keep up with thedevelopment of urban areas, as pro-

posed by several studies of regionaldisparities (Phillips, 1978). Instead, itassumes that the existence of rural-urban disparities is created byregional development policies, espe-cially for the regional componentsthat can be affected by the process ofregional development and its under-lying policies. The assumption isbased on the fact that regional poli-cies applied in the NCR favor urbandevelopment.

This study accepts the fact thatdistributed regional componentsresulting from developmentcontribute to the welfare of theregion’s inhabitants withoutentering into the debate whetherthose components are subjective orobjective measures of rural-urbandisparities. In the literature ofregional disparities, assertions aboutthe components distributed are stillnot conclusive. Some usedistribution of income per capita,wealth, or a combination of incomeand wealth as a measure of regionaldisparities (Smith, 1982). Others useand develop indicators of well-being, quality of life or standards ofliving as a measure of regional dis-parities (Knox, 1974; Kuz, 1978;Schultink, 1992). According toMatthews (1983), the meanings ofthose equity or quality-of-lifemeasures remain vague and seem tooverlap with few clear prescriptions

Number of industries in Mean of SignificantTypes of industry the village paired differences

1982 1992 %-Changedifferences

In the NCR :

Large-scale industry (LSI) 0.0909 4.4091 4750.50 -4.3182 *****

Medium-scale industry (MSI) 0.7273 2.6136 259.36 -1.8864 ***

Small-scale industry (SSI) 6.0909 11.0455 81.34 -4.9545 ****

LSI /1,000 village population 0.0177 0.7709 4255.37 -0.7532 *****

MSI/1,000 village population 0.1343 0.4052 201.71 -0.2709 ***

SSI/1,000 village population 0.9514 2.1286 123.74 -1.1773 ****

Table 16. Differences in regional industrialization between 1982 and 1992.

Source: primary data (village survey)Significance level :* 20 percent ** 15 percent *** 10 percent **** 5 percent ***** 1 percent

36

of what is involved in anyassessment of them. Therefore, thisstudy uses only the available censusdata of 1990 as measures of rural-urban disparities. This informationincludes some aspects of economicprogress, including improvements ofmarketing facilities, educationalfacilities and health care facilitiesthat may result from regional devel-opment and affect the well-being ofthe region’s inhabitants.

The distribution of those measuresis assessed using the Ginicoefficients (GCs) and the Lorenzcurves. The Gini coefficient is a sys-tem measure that can describe themagnitude of the distribution withonly one coefficient, rangingbetween 0 and 1. A value close to 0indicates equal distribution and avalue close to 1 indicates unequaldistribution. The distribution canalso be seen in the Lorenz curve, inwhich the degree of inequality isdepicted by how far the Lorenzcurve departs from the diagonal orequal-share line (Smith, 1982). Theproportion of the total area belowthe diagonal that is above the Lorenzcurve is also a measure of the GC.The GC and the Lorenz curve aresystem measures; the spatial concen-trations of specific components aremeasured using location quotients(LQs). The LQs can be used to com-pare the concentration of the compo-nents in each subregion of the NCR— less than 1 indicates that the sub-region’s share in the components inquestion is less than the region’sshare; more than 1 indicates that thesubregion’s share is higher than thatof the whole region.

The data used to analyze rural-urban disparities are 1990 censusdata. The analysis addresses theresults of the development processas reflected in the spatialdistribution of public benefits.Because of limited data availability,the study did not consider distribu-tion of development benefits amongsocial strata.

Regional Economic ProgressSeveral studies in the NCR

indicate that significant progress hasoccurred in the economy of theregion. This progress is indicated bythe increased relative share of indus-tries and services, with theirincreased value-added contributionsto the regional economy (Hill, 1992b;Soemarwoto, 1992). In addition, thisstudy also indicates that in the past10 years significant increasesoccurred in the number of industriesof all scales, the percentage of thepopulation engaged in the manufac-turing sector, the numbers ofeconomic institutions (such asbanks), and improvements in publictransportation and its infrastructure,educational facilities and health carefacilities. All of the increases indicatea growing regional economy.

Because regional developmentpolicies have favored urbandevelopment, it is expected that thebenefits of a growing regional econo-my are concentrated in urban areasand a small portion of the region’sinhabitants. In this study, the bene-fits of a growing regional economyare indicated by the increased num-ber of households with telephones,electricity or television—indicatorsof increased household economicwell-being. These indicators can beused to assess the spatialdistribution of benefits of regionaleconomic development. The resultsof the analysis are summarized inTable 22 and Figure 8.

Table 17 summarizes: (i) thelocation quotients (LQs) of the house-holds with telephones, householdswith electricity provided by PLN(Perusahaan Listrik Negara, a monop-olistic agency providing electricity tothe nation) and households with tele-vision for each subregions of the NCR;(ii) the Gini coefficients (GCs) of thoseindicators in rural areas, urban areasand the NCR; (iii) the distribution(percentage) of the indicators andpopulation in each subregions of theNCR; and (iv) the cumulative percent-ages of the indicators and populationthat indicate concentration or distribu-tion in the region.

Figure 9 presents Lorenz curvesshowing the inequality in the distri-bution of the NCR households withtelephones, electricity or televisionin relation to the NCR population.

The Gini coefficient of the house-holds with telephones in the NCR is0.77, indicating that households withtelephones are unequally distributedin the region. This high value of theGC is also represented by thedistance between the Lorenz curveand the equal-share line.

The location quotient values showthat the households with telephonesare concentrated in the most urban-ized subregions, U3 and U4, withvalues of 1.95 and 8.21, respectively.Most of the households withtelephones are located in the mosturbanized area (U4). U4 has only10.39 percent of the NCR populationbut 85.27 percent of the telephonesin the region.

Households with telephones seemvery unevenly distributed. Theurban areas have 22.23 percent of theNCR population and 91.98 percentof the telephones in the region. Onthe other hand, the 77.77 percent ofthe NCR population that lives in therural areas has only 8.02 percent ofthe telephones.

Within both rural and urban areas,the households with telephones arealso unequally distributed. This isindicated by high values of Ginicoefficients of the households withtelephones in both rural and urbanareas — 0.41 and 0.46, respectively.Those indicate that only the elitepopulation in rural areas enjoys thisconvenience.

The distribution of householdswith electricity is similar to the dis-tribution of households withtelephones and is more concentratedin urban areas, though theconcentration is less than that ofhouseholds with telephones. TheGini coefficient of the householdswith electricity in the region is 0.30(compared with households withtelephones, 0.77). Nevertheless, thisGC value also indicates that thehouseholds with electricity areunequally distributed in the region.

Region Location quotients Percentage Cumulative(LQ) distribution percentages *)

Phone Elect TV Phone Elect TV Pop Phone Pop Elect Pop TV Pop

R1 0.50 0.17 0.26 3.86 1.33 1.97 7.68 85.27 10.39 33.41 10.39 58.22 10.39

R2 0.00 0.41 0.32 0.04 11.16 8.67 27.06 89.31 12.46 39.01 12.46 62.84 12.46

R3 0.08 0.74 0.38 2.37 21.27 11.02 28.76 93.17 20.14 43.26 15.04 66.28 15.04

R4 0.12 0.98 0.55 1.75 13.92 7.80 14.27 95.19 27.33 52.32 22.23 70.54 22.23

U1 0.28 1.26 0.59 2.02 9.06 4.26 7.19 95.84 29.91 66.24 26.50 78.34 26.50

U2 0.25 1.65 1.33 0.65 4.25 3.44 2.58 97.59 44.18 87.51 65.26 89.36 65.26

U3 1.95 2.71 2.23 4.04 5.60 4.62 2.07 99.96 72.94 98.67 92.32 98.03 92.32

U4 8.21 3.22 5.60 85.27 33.41 58.22 10.39 100 100 100 100 100 100

Gini coefficient Source of data: population census of 1990.

(GC)

NCR 0.77 0.30 0.51

Rural 0.41 0.18 0.09

Urban 0.46 0.18 0.04

37

This unequal distribution of house-holds with electricity can also beseen in the Lorenz curve of Figure 8.

The values of the LQs show thatthe households with electricity aremore concentrated in urban areas —with values between 1.26 in U1 and3.22 in U4 — than in rural areas —with values between 0.17 in R1 and0.98 in R4. Region U4, with 10.39percent of the regional population,has 33.41 percent of this utility pro-vided by the government.Comparing rural and urbanconditions shows that urban areas —with 22.23 percent of the NCR popu-lation — have 52.32 percent of theelectricity, while rural areas — with77.77 percent of the population —have only 47.68 percent of this con-venience. Within rural and urbanareas, on the other hand, householdswith electricity seem equally distrib-uted with the same GC value.

With the advanced telecommuni-cation technology of satellites andwith the capability to cover allIndonesian islands, it is expected

that television wouldbe found relativelyequally distributed inIndonesia. This studyfinds that that iscertainly not the case—there is disparity inthe regional distribu-tion of householdswith television in theNCR, the most devel-oped region inIndonesia.

The Gini coefficientof the households withtelevision in the NCR is0.51, indicatingunequal distribution.This unequal distribu-tion can also be seen inthe Lorenz curve of thehouseholds with televi-sion in Figure 7.

The LQ values also show thathouseholds with TV are moreconcentrated in the most urbanizedsubregions, U2-4, with values of1.33, 2.23 and 5.60, respectively. This

indicates that most of thehouseholds with television are locat-ed in the most urbanized area (U4).U4, with only 10.39 percent of theNCR population, has 58.22 percentof the households with television.

Table 17. Location quotients (LQ), Gini coefficients (GC), percentage distribution of number of telephones (phone), houses with electricity (elect), television (TV) and population (pop.) by village; and data for Lorenz curves in the NCR, 1990.

*) Cumulative percentages are calculated based on the rank of their location quotient values and are used to construct Lorenz curves.

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100

Cum. % of Aspects of Ec. Progress

Cum. % NCR Population

equal-share

telephone

electricity

television

Cum

. % N

CR

Pop

ulat

ion

Figure 7. Lorenz curves for the distribution of the households with telephones, electricity andtelevision in relation to population, by subregions of the NCR (source: Table 22).

38

The distribution of the householdswith television between rural andurban areas also seems veryunequal. Urban areas, with only22.23 percent of the NCR population,have 70.54 percent of the total televi-sions in the region.

The unequal distribution of thethree indicators shows that the bene-fits of the growing regional economyof the NCR are enjoyed mostly bythe urban population and by a smallportion of the rural population.

Marketing FacilitiesThe study assessed the

distribution of regional marketingfacilities in the NCR, specifically, thedistribution of marketing facilitiesfor agricultural inputs, marketingfacilities for agricultural productsand local markets with permanentbuildings.

The NCR is the predominant agri-cultural region in the country, so itwould be expected that marketingfacilities for agricultural inputs and

outputs would be located in ruralareas or equally distributed through-out the region. This study uses thenumber of kiosks selling agriculturalinputs (KAI) as the indicator of theavailability of marketing facilities foragricultural inputs, and the numberof kiosks selling agriculturalproducts (KAP) as the indicator ofthe availability of marketingfacilities for agricultural products.All local markets are developed andoperated by the government. Thepresence of a market withpermanent buildings is used as anindicator of the degree ofgovernment commitment to developthe local economies.

The results of the analysis aresummarized in Table 18 and Figure9. Table 18 summarizes: (i) the LQsof the markets with permanentbuildings (market), kiosks sellingagricultural inputs (KAI) and kiosksselling agricultural outputs (KAP)for each subregion of the NCR; (ii)the GCs of those indicators in ruralareas, urban areas and the NCR; (iii)

percentage distribution of the indica-tors and population in eachsubregion of the NCR; and (iv)cumulative percentages of the indi-cators and population in the NCR.

Figure 9 presents Lorenz curvesshowing the inequality in the distri-bution of markets with permanentbuildings (market), kiosks sellingagricultural inputs (KAI) and kiosksselling agricultural products (KAP)in relation to the NCR population.

As expected, the results of theanalysis show that kiosks sellingagricultural inputs (KAI) and thoseselling agricultural products areevenly distributed in the region andin the rural areas. The Ginicoefficients of the distribution ofKAI in the region and in rural areasare 0.04 and 0.03, respectively. TheGini coefficients of KAP in theregion and in rural areas are 0.15 and0.08, respectively. These indicate thatboth kiosks selling agriculturalinputs (KAI) and kiosks selling agri-cultural products (KAP) are equallydistributed in the region.

Region Location quotients Percentage Cumulative(LQ) distribution percentages *)

Market KAI KAP Market KAI KAP Pop Market Pop KAI Pop KAP Pop

R1 0.42 0.86 0.71 3.20 6.64 5.46 7.68 28.80 10.39 2.93 2.58 6.46 2.58

R2 0.86 0.98 0.75 23.20 26.58 20.20 27.06 34.40 12.97 5.28 4.65 28.28 12.97

R3 0.56 1.07 1.00 16.00 30.75 28.89 28.76 37.60 15.04 20.85 18.92 57.17 41.73

R4 0.67 1.09 0.72 9.60 15.57 10.30 14.27 48.00 22.23 51.60 47.68 62.83 48.92

U1 1.45 1.02 0.79 10.40 7.36 5.66 7.19 71.20 49.29 58.96 54.87 83.03 75.98

U2 2.17 1.14 2.50 5.60 2.93 6.46 2.58 80.80 63.56 85.54 81.93 93.33 90.25

U3 1.55 1.14 0.58 3.20 2.35 1.21 2.07 96.80 92.32 92.18 89.61 98.79 97.93

U4 2.77 0.75 2.10 28.80 7.82 21.82 10.39 100 100 100 100 100 100

Gini coefficient Source of data: population census of 1990.

(GC)

NCR 0.26 0.04 0.15

Rural 0.10 0.03 0.08

Urban 0.13 0.09 0.22

Table 18. Location quotients (LQ), Gini coefficients (GC), percentage distribution of village marketing facilities, including number of markets with permanent buildings (Market), kiosks selling agricultural inputs (KAI), kiosks selling agricultural products (KAP) and population (Pop.); and data for Lorenz curves in the northern coastal region (NCR) of West Java, 1990.

*) Cumulative percentages are calculated based on the rank of their location quotient values and are used to construct Lorenz curves.

39

Especially for input kiosks, theLorenz curve shows that the line isclose to the equal-share line (Figure9), indicating that spatialdistribution is nearly equal. This isalso supported by the values of LQs,which are essentially identical for allsubregions of the NCR.

However, the Gini coefficient ofthe markets with permanent build-ings in the region (0.26) is higherthan the other two indicators. Thisindicates that spatial inequalityexists in the distribution of govern-ment markets with buildings, asindicated by the large area betweenthe Lorenz curve and the equal-share line of Figure 8.

The values of LQs show that mar-kets with permanent buildings aremore concentrated in urban areas —with values between 1.45 and 2.77 —compared with rural areas, with val-ues between 0.42 and 0.67. Urbanareas have 22.23 percent of the NCRpopulation but 48 percent of market-ing facilities, while rural areas with77.77 percent of the population haveonly 52 percent of the facilities.

The distribution of the three indi-cators shows that marketingfacilities developed by thegovernment (markets withpermanent buildings) tend to be

concentrated in urban areas, whilemarketing facilities developedthrough private initiatives (KAI andKAP) tend to be more equallydistributed in the NCR.

Educational FacilitiesThis section describes the distrib-

ution of pre-college educationalfacilities in the NCR, including ele-mentary schools, secondary schoolsand high schools. The results of theanalysis are presented in Table 19and Figure 10.

Table 19 summarizes: (i) the loca-tion quotients (LQs) of elementaryschools (Elsch), secondary schools(Secsch) and high schools (Hisch)for each subregion of the NCR; (ii)the Gini coefficients (GCs) of theindicators in rural areas, urban areasand the NCR; (iii) percentage distri-bution of the indicators and popula-tion in each subregion of the NCR;and (iv) cumulative percentages ofthe indicators and population in theNCR. Figure 10 presents Lorenzcurves showing the inequality in thedistribution of elementary schools,secondary schools and high schoolsin relation to the NCR population.

The Gini coefficient of elementaryschools in the NCR is 0.04, indicat-

ing that elementary schools areevenly distributed in the region. ItsLorenz curve is close to the equal-share line of Figure 10.

Although all values of the LQs inthe urban areas are higher thanthose in rural areas, the differencesare very small, indicating thatelementary schools are evenly dis-tributed in all subregions of theNCR. Within rural and urban areas,the GC values are 0.01 and 0.05,respectively.

The distribution of secondaryschools is more concentrated inurban areas, with a regional GCvalue of 0.23. This is also reflectedby the Lorenz curve (Figure 10).

The location quotients show thatsecondary schools are the most con-centrated in the most urbanized area(for U4 — 2.84, U3 — 1.56 and U1 —1.43.

The Gini coefficient of highschools in the NCR is 0.48, indicat-ing an uneven distribution of highschools in the region, alsodemonstrated by the Lorenz curvesof high schools in Figure 10.

The values of LQs show that highschools are clearly most concentrat-ed in the urban areas (U1-4), withvalues ranging between 1.22 in U2and 4.83 in U4. The share of theregional population for urban areasis only 22.23 percent, but these areashave 69.95 percent of the highschools. In addition, high schoolsamong urban subregions are notevenly distributed, as indicated bythe GC value of .25.

The distribution of the three indi-cators of educational developmentin the NCR shows a clear hierarchi-cal distribution of public education-al facilities. This distribution reflectsan educational development policywith basic educational facilitiesequally distributed in the regionand a concentration of secondaryeducational facilities in severalregional centers. Rural people whowish to pursue higher education areexpected to go to the regionalcenters.

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100

Cum. % of Marketing Facilities

Cum. % of NCR Population

equal-share

KAI

KAP

Market

Cum

. % o

f NC

R P

opul

atio

n

Figure 8. Lorenz curves for the distribution of markets with permanent building (Market), kiosks selling agricultural inputs (KAI), and kiosks selling agricultural products (KAP) in relation to population, by subregions of the NCR (source: Table 23).

40

Region Location quotients Percentage Cumulative(LQ) distribution percentages *)

Elsch Secsch Hisch Elsch Secsch Hisch Pop Elsch Pop Secsch Pop Hisch Pop

R1 0.98 0.39 0.18 7.50 3.04 1.35 7.68 2.71 2.07 29.49 10.39 50.22 10.39

R2 0.95 0.63 0.38 25.64 16.97 10.31 27.06 15.85 12.46 32.72 12.46 56.50 12.46

R3 0.96 0.82 0.42 27.68 23.64 12.11 28.76 19.01 15.04 43.02 19.65 66.81 19.65

R4 0.91 0.78 0.44 13.05 11.11 6.28 14.27 26.13 22.23 45.24 22.23 69.95 22.23

U1 0.99 1.43 1.43 7.12 10.30 10.31 7.19 33.63 29.91 68.88 50.99 76.23 36.50

U2 1.22 1.22 1.22 3.16 2.22 3.14 2.58 61.31 58.67 79.99 65.26 88.34 65.26

U3 1.31 1.56 3.03 2.71 3.23 6.28 2.07 86.95 85.73 96.96 92.32 98.65 92.32

U4 1.26 4.83 4.83 13.14 29.49 50.22 10.39 100 100 100 100 100 100

Gini coefficient Source of data: population census of 1990.

(GC)

NCR 0.04 0.23 0.48

Rural 0.01 0.08 0.06

Urban 0.05 0.18 0.25

Table 19. Location quotients (LQ), Gini coefficients (GC), percentage distribution of village educational facilities including number of elementary schools (Elsch), secondary schools (Secsch), high schools (Hisch) and population (Pop.), and data for Lorenz curves in the Northern Coastal Region of West Java, 1990.

*) Cumulative percentages are calculated based on the rank of their location quotient values and are used to construct Lorenz curves.

Health Care FacilitiesThis section describes the distribu-

tion of health care in the NCR,including physicians, traditional mid-wives and health care facilities —hospitals, polyclinics, clinics, materni-ty centers and community health carecenters. It is important to note thattraditional midwives in the regionprovided extended services to localcommunities, including the generaltreatment of illnesses. The results ofthe analysis are summarized in Table20 and Figure 10.

Table 20 summarizes: (i) the loca-tion quotients (LQs) of the numbersof physicians, traditional midwives(mwife) and total health careproviders (care) for each subregionof the NCR; (ii) the Gini coefficients(GCs) of the indicators in rural areas,urban areas and the NCR; (iii)percentage distribution of the indica-tors and population in eachsubregion of the NCR; and (iv)cumulative percentages of the indi-cators and population in the NCR.Figure 11 presents Lorenz curves

showing inequality inthe distribution ofphysicians, traditionalmidwives and healthcare providers in rela-tion to the NCR popu-lation.

The Gini coefficientof physicians in theNCR is 0.63, indicatingthat physicians areunevenly distributedover the region.Physicians in theregion are concentrat-ed in the most urban-ized region (U4) withan LQ of 6.74, followedby U2 with 2.43. Theurban areas (U1-4)have only 22.23 percent of theregional population but 80.50percent of physicians. Within urbanareas, the physicians are alsounevenly distributed among thesubregions (GC – 0.40).

Physicians represent moderntreatments in health care andmidwives represent traditional care

in childbirth. LQs show thattraditional midwives are evenly dis-tributed over space, indicating thattraditional treatments in health careare still popular in both rural andurban areas of the NCR. Table 20and Figure 11 show that traditionalmidwives are evenly distributed

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100

Cumulative Percent of School

Cumulative Percent of NCR Population

equal share

Elsch

Secsch

Hisch

Cum

ulat

ive

Perc

ent o

f NC

R P

opul

atio

n

Figure 9. Lorenz curves for the distribution of elementary schools (Elsch), secondary schools (Secsch) and high schools (Hisch) in relation to population, by subregions of the NCR (source: Table 19).

41

Region Location quotients Percentage Cumulative(LQ) distribution percentages *)

Phys Mwife Care Phys Mwife Care Pop Phys Pop Mwife Pop Care Pop

R1 0.30 1.00 0.82 2.31 7.67 6.26 7.68 70.08 10.39 3.42 2.58 20.97 10.39

R2 0.16 1.06 0.86 4.25 28.76 23.19 27.06 75.10 12.46 5.97 4.65 23.48 12.46

R3 0.22 0.94 0.84 5.02 21.63 19.50 28.76 79.15 19.65 34.73 31.71 26.54 15.04

R4 0.40 0.98 0.92 7.92 19.44 18.33 14.27 80.50 22.23 42.40 39.39 44.87 34.97

U1 0.56 0.86 0.86 4.05 6.18 6.18 7.19 88.42 42.16 52.75 49.78 68.06 62.03

U2 0.52 1.33 1.19 1.35 3.42 3.06 2.58 90.73 49.84 72.19 69.71 74.24 69.22

U3 2.43 1.23 1.21 5.02 2.55 2.51 2.07 95.75 72.94 93.82 92.81 93.74 92.32

U4 6.74 1.00 2.02 70.08 10.35 20.97 10.39 100 100 100 100 100 100

Gini coefficient Source of data: population census of 1990.

(GC)

NCR 0.63 0.03 0.12

Rural 0.17 0.02 0.02

Urban 0.40 0.20 0.17

Table 20. Location quotients (LQ); Gini coefficients (GC), percentage distribution of village availability of health care facilities including numbers of physician (Phys), traditional midwives (Mwife), health care providers (Care) and population (Pop.); and data for Lorenz curves for the Northern Coastal Region of West Java, 1990.

*) Cumulative percentages are calculated based on the rank of their location quotient values and are used to construct Lorenz curves.

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100

Cum % of Health Care Related Facilities

Cum. % of NCR Population

equal-share

Doctors

Midwives

Care

Cum

. % o

f N

CR

Pop

ulat

ion

Figure 10. Lorenz curves for the distribution of the number of physicians, midwives and total health care providers (Care) in relation to population, by subregions of the NCR (source: Table 20).

over all subregions of the NCR witha GC value of 0.03.

The Gini coefficient of total healthcare providers in the NCR is 0.12,indicating that the health careproviders are evenly distributedover space. However, the LQ valuesare still higher in urban areas than inrural areas, indicating that urbanareas enjoyed a larger share of healthcare providers.

42

Summary and PolicyImplications

Urbanization in the NorthernCoastal Region (NCR) of West Java,a corridor region connecting Jakartaand Cirebon and the major food(rice) production region ofIndonesia, exemplifies a public poli-cy challenge for Indonesian agricul-tural and regional development. Theregion contains the best irrigationsystems and agriculturaldevelopment institutions inIndonesia, has long been the centerof agricultural development and is,therefore, considered a nationalmodel. At the same time, the regionis experiencing massive land conver-sion pressure because of regionalindustrialization and developmentpolicies favoring urbanization.

Historical land use practices andcontemporary development policieshave helped to shape the agriculturein the region. At the national level,the GOI’s goal has been to achieveand maintain Indonesia’s rice self-sufficiency, which was accomplishedin the mid-1980s. Consequently, thestatus of the NCR as a major foodproducer in Indonesia will likely bemaintained. This reality is alsoreflected in current agricultural poli-cies such as price supports for agri-cultural commodities, subsidies forthe development and rehabilitationof irrigated paddylands, theprovision of agricultural inputs andother supply measures to counteractsignificant fluctuations in agricultur-al commodities prices. At the sametime, by presidential decree and stat-ed regional planning objectives, theGOI aims to restrict conversion ofagricultural land (especially irrigat-ed paddyland) into non-agriculturaluses. This policy reflects concernsabout urbanization impacts.

In reality, however, developmentalimpacts in the NCR seem to contra-dict the GOI’s goal to maintain theregion as the “rice bowl” ofIndonesia. Regional development ofthe NCR is part of the developmentof the larger Jakarta-West Javaregion, designated as the hinterland

of Jakarta. More importantly, indus-trial and commercial sector develop-ment, which is deemed saturated inthe Jakarta metropolitan area, seemsto reflect public development policytargeted for West Java. The reality isthat development favors areas with abetter infrastructure and proximityto government institutions. Becausethe NCR has both a better infrastruc-ture and proximity to Jakarta, indus-trial development in West Java hasbecome more concentrated in thisregion.

The regional development plan ofWest Java adopts functional integra-tion rather than a territorial integra-tion as its strategy of regional plan-ning, with “growth centers” as theprincipal instrument of its spatialdevelopment policy. The growthcenter policy is to induce urbaniza-tion through a combination of directpublic investments and capital sub-sidies to private enterprises (or theprivate sector, at large). In contrast,functional integration is a strategybased on the mutual self-interestsamong regions. Given inequalities atthe start, functional integration ischaracterized by the inequalities of ahierarchical power structure, whileterritorial relationships are temperedby mutual rights and obligationsamong participating jurisdictions.

A regional development strategybased on functional integration isbiased toward urban developmentand characterized by largeinvestment in the urban-industrialsector. Two important problems areassociated with urbanization drivenby such a regional developmentstrategy:

1. Agricultural development will benegatively affected by urbaniza-tion because development ofurban sectors is viewed as themajor determinant of regionaleconomic growth.

2. Rural areas take part in the gen-eral process of growth diffusiononly to the extent that they aresubject to the impacts of theurban economy. In other words,the urbanization process is notnecessarily associated with social

urbanization (e.g., a wider distri-bution of social services andsocioeconomic benefits) andincreasing opportunities fordevelopment in rural areas.Therefore, it is expected thatrural-urban disparities in the dis-tribution of the developmentbenefits result from a cumulativecausation process or self-reinforc-ing cycle.

The general objective of this studyhas been to provide a better under-standing of the urbanization processand its causal factors, itsrelationships with regional develop-ment and most importantly, itsimpacts on agricultural developmentin the region. This objective address-es the goal to identify a rationalframework for identifying and eval-uating policy interventions tosustain a viable agricultural sectorand promote comprehensive andsustainable development. In summa-ry, this study addressed the follow-ing objectives :

1. An analysis of urbanization inthe NCR and the identification ofassociated factors such aschanges in land tenure, the agri-cultural economy, demographicsand educational opportunities;the development in public facili-ties and industrial development;and proximity to primary cities.

2. A description of the impacts ofurbanization on agriculturaldevelopment, especially landtenure and the structure oflandownership; the degree ofabsentee landownership; agricul-tural prime land conversion andits subsequent impact on agricul-tural household income; and therural and regional employmentstructure.

3. A description of the spatialimpacts of urbanization on theagricultural resource base, thedistribution of public facilities,and the disparities in economicdevelopment among rural andurban areas within the region.

4. A description of the functionalrelationship among urbanization,

43

regional planning and agricultur-al development.

5. An examination of policy impli-cations for sustainable agricultur-al and overall development plan-ning in the NCR.

The following section summarizesthe research findings, presentspotential policy implications of thesefindings and identifies needs for fur-ther research.

Summary of FindingsThis research yielded findings in

five important areas: (i) the regionalurbanization process and its under-lying causal factors; (ii) the level ofregional urbanization; (iii) theimpacts of urbanization on regionalagricultural development, especiallyon the structure of landownership,the degree of absenteelandownership, conversion of agri-cultural (prime) land and its impacton the agricultural householdincome, and the structure of ruralemployment; (iv) regional industri-alization; and (v) rural-urban dispar-ities.

1. The process of urbanization.

Regional urbanization during theperiod 1982-1992, as measured byrate of change in population density,is caused by three groups of factors:(i) rural factors or push factors, (ii)urban factors or pull factors, and (iii)demographic factors. Multiple regres-sion analysis indicates that 76.84 per-cent of urbanization can beexplained by those three groups ofcausal factors. From this total, ruralfactors contribute 45.40 percent (59.08percent of the explainedurbanization). Urban factorscontribute 24.72 percent (32.17 per-cent of the explained urbanization)and demographic factors contribute6.72 percent (8.75 percent of theexplained urbanization).

Among rural factors, the increasein agricultural land parcelization (i.e.,increased landless households anddecreased households with agricul-tural lands) contribute 28.16 percentor 36.65 percent of the explained

urbanization or 62.03 percent of therural factors. The contribution of theagricultural land parcelization inexplaining regional urbanization ishigher than all pull factors (24.72 per-cent). Changes in the agricultural econ-omy (i.e., decreased real agriculturallabor wage, increased rural landprices and changes in the villagestructure of employment) contribute17.24 percent or 22.44 percent of theexplained urbanization or 39.97 per-cent of rural factors. Increases inagricultural land conversion contribute12.38 percent or 16.11 percent of theexplained urbanization or 27.27 per-cent of rural factors. Finally, increasesin the absentee landownershipcontribute 8.62 percent or 11.22 per-cent or 18.99 percent of rural factors.

Among urban factors, increases inthe availability of public facilities explainmost of the regional urbanization,followed by increases in the availabil-ity of pre-college educationalfacilities. Surprisingly, besides thefact of increased regional industrializa-tion, this factor appears insignificant inaffecting regional urbanization. Thetotal contribution of demographic fac-tors explaining regional urbanization,meanwhile, is only 6.72 percent.

In summary, findings show thatregional urbanization is dictated mainlyby the worsening structural conditionsin agricultural land tenure and a deteri-oration of socioeconomic conditions inrural areas compared with the moreurbanized subregions.

Regarding the proximity ofvillages to the cities, results showthat village proximity to Cirebonand to the capital cities ofkabupatens does not significantlyaffect regional urbanization.Proximity of the villages to Jakarta,meanwhile, affects regionalurbanization (confidence level of 70percent), as does proximity ofvillages in urban areas (urbanvillages) to Jakarta (confidence levelof 78 percent). Moreover, urbanareas affect urbanization (confidencelevel of 95 percent), an indicationthat urbanization as a regionalprocess is accelerated through a self-reinforcing cycle.

2. Level of regional urbanization

The level of urbanization isdefined at the desa or village levelby categorizing villages into ruralvillages and urban villages. Urbanpopulation is defined as the popula-tion of urban villages.

Using a population density of1,500 people/km2 to differentiatebetween rural and urban villages,the study finds that during the peri-od 1980-1990, the number of urbanvillages in the region increased from151 villages to 223 villages or from13.74 percent to 20.29 percent, a sub-regional rural village urbanization rateof 4.77 percent per year. Similarly, theurban population increased from1,179,362 people to 2,439,996 peopleor from 25.69 percent to 38.75percent, a regional urbanization rate ofpercent or 10.69 percent per year.

The analysis of temporal rate ofchange in population density(TRCPD), furthermore, shows thatpopulation density increasedthroughout the region. However,this rate decreases comparativelyfrom the most urbanized area to themost rural area, illustrating a contin-ued concentration of population densi-ties. The mean of TRCPD is 115.57people/km2/year in urban areas andin rural areas 12.76 people/km2/year, indicating that the annualincrease in population density in the“urban” areas is 9.01 times greater thanin “rural” areas. This may indicate apreference to locate in more urbanizedareas because of the perceived higherquality of life there.

The analysis of spatial rate ofchange in population density(SRCPD), a difference in the degreeof population concentration overdistance from the capital cities of thekabupatens, shows that the values ofSRCPD are very low in allkabupatens in the region. TheSRCPD in the NCR, Bekasi,Karawang, Subang and Indramayuare 0.0441, 0.0443, 0.0485, 0.0440 and0.0417, respectively. This means thatin Bekasi, for example, for each kilo-meter farther from the capital city ofthe kabupaten, the population densi-ty declines by only 0.0443 people per

44

square kilometer. Therefore, nosignificant difference exists inpopulation density over distance fromthe capital city of the kabupaten.Therefore, a relatively even spatialdistribution of population in thekabupatens is present and the ranksize rule does not represent the prob-able distribution of population in theNCR. In other words, the simplenegative exponential function,which represents a successful expla-nation model for the developedcountries, is not representative forthe NCR.

3. Impacts of urbanization on agricultural developmentThe analysis shows that a high

annual regional urbanization rate isaccompanied by an increase in landparcelization characterized by anincrease in landless households, anincrease in absentee landownership, anincrease in agricultural to non-agricul-tural land conversion, and structuralchanges in village employment, both inrural areas and in the NCR as awhole. The impacts are, however,more dominating in rural areas.

At the regional level, the analysisshows that the percentage oflandless households increased dur-ing the 1982-92 period. Householdswithout agricultural drylandincreased by 18.84 percent and thosewithout paddyland increased by22.84 percent. The increase in region-al landless households is not signifi-cantly associated with a decrease inthe percentage of households with <0.5 ha. dryland or < 0.5 ha.paddyland. This indicates that thepercentage of subsistence farmers inthe region has remained largely thesame. Regional agriculturaldevelopment has, therefore, been unableto reduce the percentage of subsistencefarmers. However, the increase ofregional landless households is associat-ed very significantly with the decreasesin the percentage of households with >0.5 ha. dryland or paddyland.

In the rural areas, changes in agri-cultural landownership are similarto those in the NCR as a whole.Landless households increased and thehouseholds with agricultural landhold-

ings decreased. The percentage ofhouseholds without agriculturaldrylands and paddylands increasedvery significantly by 16.95 percentand 17.82 percent, respectively. Thepercentage of households with > 2.0ha dryland and those with the samehectarage of paddyland decreasedby 58.04 percent and 49.56 percent,respectively.

The analysis also shows that theurbanization is accompanied byincreased absentee landownership.At the regional level, the total absen-tee lands and the number of absen-tee landowners increased verysignificantly — by 193.20 percentand 43.26 percent, annual increasesof 19.30 percent and 4.33 percent,respectively. The total hectarage ofabsentee lands increased faster than thenumber of absentee landlords, indicatinga concentration of absentee landowner-ship.

The increase in absentee land areacan be attributed mostly to anincrease in absentee dryland rather thanto an increase in absentee paddyland.Absentee dryland in the regionincreased by 605.94 percent duringthe period 1982-92, or 60.59 percentannually. Absentee paddyland,meanwhile, increased 42.29 percentduring the same period, or 4.23 per-cent annually.

This pattern of increase in absen-tee dryland is accompanied by ahigher number of absentee landown-ers acquiring dryland than acquiringpaddyland. The average number ofabsentee landowners with drylandincreased by 22.93 percent annually;those with paddyland increased byonly 1.85 percent annually. It is moreeconomical to invest in paddylandthan in dryland. The high increase ofabsentee dryland in the NCR might beattributed to increased regional landspeculation. Buyers usually haveinformation on rural land designat-ed for urban use, which is rarely irri-gated farmland. This is especiallyimportant, given that existing poli-cies prohibit the conversion of pad-dyland to non-agricultural uses.

In rural areas, changes in absenteelandownership are more staggering than

those in the region as a whole. The totalabsentee lands and the number ofabsentee landowners during 1982-1992 increased by 193.20 percent and43.26 percent, respectively.

Comparing the total hectarage ofabsentee lands (dryland and paddy-land), the analysis shows that absen-tee dryland increased by 638.06 percent,or 63.81 percent annually. Absenteepaddyland, meanwhile, increased by29.78, or 2.98 percent annually. Thispattern is also accompanied by ahigher number of absentee landown-ers acquiring dryland rather thanpaddyland. The number of absenteelandowners with dryland increased by46.02 percent annually, while those withpaddyland did not increase significantly,indicating that the absentee landowner-ship of paddyland in rural areas isbecoming more concentrated.

The analysis of agricultural landconversion and the urbanizationprocess in the region shows that, atthe regional level, irrigated landdecreased by 25.62 percent duringthe period 1982-1992, or 2.56 percentannually, while agricultural drylanddecreased by 27.11 percent, or 2.71percent annually. In rural areas, irri-gated land decreased by 2.50percent annually, while the agricul-tural dryland did not decrease sig-nificantly. This indicates that agricul-tural land conversion in rural areasinvolves mostly irrigated land, themainstay of regional agriculturalproduction.

At the regional level, the increasedconversion of irrigated land puts addi-tional pressure on the NCR’s capabilityto maintain its key role in Indonesianfood production. At the householdlevel, however, this study finds thatthose involved in land conversion arebetter off economically. Over the 10-year period, total household incomesincreased by 37.58 percent whenland was converted. This study alsofinds that land conversion in the NCRis mostly dictated by market prices andthat government policy does not signifi-cantly affect household decisions regard-ing land conversions.

The study finds that five fundamen-tal differences exist between rural and

45

regional structural changes of employ-ment. First, in the period of 1982-1992, rural employment appears stilldominated by agricultural employment.Second, the percentage of agriculturallaborers in rural areas increased signifi-cantly — by 22.04 percent. Thisincrease is much higher than that ofthe region (0.95 percent). Third,although employment in the manu-facturing sector in rural areasincreased relatively very significant-ly (147.73 percent) and the increaseis higher than that of the region(87.62 percent), the absolute changein manufacturing employment in ruralareas is much less than that of theregion. Fourth, the increase of small-scale enterprises is higher inrural areas than in the region, 65.55percent and 47.28 percent,respectively. Finally, the employmentratio in private enterprises in the ruralareas is much less than that in theregion as a whole.

These figures imply that, at theregional level: (i) a decrease inemployment in agriculture isfollowed by a significant increase inthe employment in manufacturingand small-scale enterprises; and (ii)a decrease in the percentage of self-employed farmers is not followedby an increase in employment in theagricultural sector, an indicationthat manufacturing industries andsmall-scale enterprises have thecapacity to absorb farmers displacedfrom the agricultural sector.Meanwhile, for rural areas thefigures imply that: (i) a decrease inemployment in the agricultural sec-tor is not followed by a significantincrease in employment in the man-ufacturing sectors; and (ii) adecrease in the percentage of self-employed farmers is followed by anincrease in employment as agricul-tural laborers and small-scale enter-prises, an indication that displacedfarmers in rural areas are eitherbeing employed as agriculturallaborers or entering the labor forceof small-scale enterprises.

4. Regional industrializationThe degree of regional industrial-

ization is measured by changes in

the numbers of large-scaleindustries, medium-scale industriesand small-scale industries. Large-scale industry is defined as an indus-try with 100 or more employees.Medium-scale industry is defined asan industry with 20 to 99 employees.Small-scale industry is defined as anindustry with 5 to 19 employees.This study finds that during the period1982-92, the degree of industrializationincreased very significantly. The num-ber of large-scale industriesincreased by 4750.50 percent, or475.05 percent annually. The numberof medium-scale industriesincreased by 259.36 percent duringthe period, or 25.94 percent annually.The number of small-scale industriesincreased by 81.34 percent duringthe period, or 8.13 percent annually.These figures show that the developmentof medium- and large-scale industries isa rather recent phenomenon associatedwith recent regional economic growth.

5. Rural-urban disparitiesThe study finds that rural-urban

disparities exist in the distribution ofthe benefits of regional economicprogress as measured in 1990. This isexpressed by the high values of theGini coefficients (GC) of householdswith telephones, electricity or televi-sion — 0.77, 0.30 and 0.51, respective-ly. Of the three, telephones are mostevenly distributed and rural electrifi-cation remains a significant developmentobjective. Economic progress in theregion is more concentrated in the moreurbanized areas since the locationquotients (LQ) of the three indicators arehigher than those in rural areas.

The marketing facilities associatedwith the sale of agricultural inputsand products are evenly distributedin the region, with GC of 0.04 and0.15, respectively. However, marketswith permanent buildings areunevenly distributed in the region(GC of 0.26) and most areconcentrated in urban areas.Elementary schools are evenlydistributed in the region, with a GCof 0.04 and with relatively similarLQs between rural and urban areas.Secondary schools and high schoolsare unevenly distributed in the

region, with GCs of 0.23 and 0.48,respectively, and concentrated most-ly in urban areas, as indicated bycomparatively higher LQs.

The study also finds thatphysicians are concentrated mostlyin urban areas. With only 22.23 per-cent of the regional population, theurban areas have 80.50 percent of thephysicians residing in the region(GC = 0.63). On the other hand, mid-wives and total health care providersare evenly distributed, with GCs of0.03 and 0.12, respectively.

Policy ImplicationsResearch findings suggest that

without any policy interventions inregional economic and agriculturaldevelopment, future agricultural via-bility can not be sustained and the roleof the region as the major food producerof Indonesia will be undermined.Therefore, regional urbanization isdirectly threatening Indonesia’sfuture self-sufficiency in rice. Moreimportantly, the welfare of the popu-lation engaged in agricultural activi-ties — 49.28 percent of the regionalpopulation or 75.88 percent of thetotal rural population (1992) — willalso be negatively affected by thisprocess.

These assertions are supported bythe research findings indicating thaturban and industrial development havedisplaced self-employed farmers throughincreased agricultural land conversionand increased absentee landownership.The logical social adjustmentcommonly pursued by rural popula-tions is to disperse household agri-cultural land by means of land con-version and land sales. This landparcelization results in diseconomiesof scale in agricultural enterprisesand provides additionaldisincentives for rural inhabitants toremain employed in agriculture. Thecombined effects of increased agricultur-al land conversion, increased absenteelandownership and increased landparcelization will decrease regional agri-cultural output and agriculturalproductivity. In addition, increasedregional industrialization has beenunable to absorb the population

46

displaced from the agricultural sector.Consequently, this labor force shiftsto small-scale, non-agriculturalenterprises or remains in agricultureas hired laborers. Because these lasttwo forms of employment are char-acterized by low economic produc-tivity and labor rates, the urbanizationprocess indirectly diminishes the overallwelfare of the rural population.

The research findings also suggestthat the existing urbanizationprocess will complicate problems ofregional development. Becauseregional development is pursuedthrough a growth center policy,urbanization is induced by a combi-nation of direct public investmentsand capital subsidies to privateenterprises to create urban centersthat are expected to diffuse econom-ic growth into rural areas. Theresearch findings show that theexpected trickle-down effects into ruralareas do not occur in the NCR. In fact,the analysis of rural-urban dispari-ties shows that the economic progressand development of public facilitiesremain largely concentrated in theurban areas. Moreover, urbanizationin the region is mostly determinedby push factors from agriculturalrural areas rather then increasedopportunities in urban areas. Thisimplies that, without furthergovernment intervention, future rural-urban disparities will increase andincreased economic stress in agriculturalrural areas will drive income flows intourban areas, creating further rural-urban disparities.

To sustain its critical role as the mainfood production region in Indonesia andto improve the welfare of its rural popu-lation engaged in agricultural activities,government must initiate a broad-basedrural development strategy. The strate-gy must include policies on:

1. Agricultural land consolidation.Agricultural land consolidation isdirected mainly to mitigate nega-tive effects of diseconomies of scale inagricultural enterprises, whichcommonly create disincentivesfor the rural population toremain in agriculture. It isurgently needed in the region for

two fundamental reasons: (i) cur-rent regional industrial and ser-vice sectors are still unable toabsorb the labor force displacedfrom the agricultural sector; and(ii) economies of scale in agricul-tural enterprises and its servicesupport infrastructure are need-ed to maintain and improveregional agricultural productivityand outputs. Policies of thisnature have proven very effectivein northwestern Europe (e.g., theNetherlands).

2. Minimizing agricultural landconversion. This policy should bepursued by the governmentthrough effective policymeasures instead of bylegislation that is difficult if notimpossible to enforce. Forinstance, this study shows thatland conversion decisions aredictated mainly by land pricesand, in fact, do have a beneficialeffect on farm household income.Economic motives and incentivesare paramount, and legalrestraints such as the relevantpresidential decree appeared tobe ineffective in reducing landconversion. Field observationsshow that farm householdschange the status of irrigatedland to non-irrigated land by tak-ing the land out of production fortwo years and then converting itto non-agricultural uses and soavoid legal impediments. Policymeasures that create direct economicincentives or purchase developmentrights on the basis of a comprehen-sive land use plan would likely bemore effective.

3. Minimizing absentee landowner-ship. This policy is directed tomaking agricultural enterprises inrural areas economically competitiveand reducing the economicincentives to sell land. In addition, aland use plan with permanent agri-cultural zoning designations willreduce the opportunity to realizehigher land prices based on landspeculation. Farm loan supports andtax incentives may provide incen-tives to preserve farm ownership

and reconvert absentee lands tosmall farmers.

4. Strengthening rural areas. Thisstudy shows that urbanization isaffected mostly by push factorsassociated with rural life.Therefore, to reduce urbanmigration indirectly, the physicaland service infrastructure of ruralareas must be improved to ensure acompetitive production environment(e.g., irrigation, transportation,marketing and service) and equi-table quality of life (e.g., housing,education, health care).

These policies may be pursued inthe context of an integrated ruraldevelopment scenario. For instance,the four potential policy initiativesabove can be addressed by thefollowing two scenarios:

1. Scenario I :The government still maintainsthe current regional developmentstrategy – namely a growth centerpolicy in which urbanization isinduced by a combination ofdirect public investments and cap-ital subsidies to private enterpris-es to create urban centers that areexpected to diffuse economicgrowth into rural areas.

Under this scenario, ruraldevelopment is pursued along withurban development (Rondinelli,1983). Within this framework, thenotion of linkages between rural andurban areas is viewed as very crucialfor rural development because majormarkets for agricultural surplusesare in urban centers, while publicservices are commonly located inurban areas. Instead of focusinginvestments exclusively on ruralareas or urban areas, investments arediversified over space in such amanner that secondary cities can becreated to build linkages betweenrural areas and urban centers. Inother words, the focus of thisscenario is to create generative secondary cities that are expected to:(i) relieve pressures on the largestcities; (ii) reduce regionalinequalities by diffusing the benefitsof urbanization; (iii) stimulate the

47

rural economy by providing publicservices, enlarging rural markets andfacilitating the absorption of agricul-tural surplus labor by urban centers;(iv) provide an increased regionallydecentralized administrative capaci-ty; and (v) help alleviate poverty inintermediate-sized cities (Rondinelli,1983).

If the expected roles of the genera-tive secondary cities can be targetedas a regional development strategy,the agricultural sector will developas a result of expanded local marketsfor agricultural products.

2. Scenario II:The government still maintainscurrent regional developmentstrategies but gives greater auton-omy to local authorities to createland use planning initiatives(Scenario I + greater local autonomy).

Under this scenario, scenario I isbroadened to include increased localautonomy to determine local landuse designations at subregional andlocal levels. To accomplish this, theregional local government mustdevelop (sub)regional and local landuse plans, incentives and controls toeffectively address, directly andindirectly, the various factors affect-ing regional agricultural sustainabili-ty, as identified in this study. Thisconstitutes policy measures under-taken in addition to thedevelopment of generative,secondary cities. Planning mustreflect regional agricultural develop-ment goals and current and futurechallenges in maintaining theregion’s position as the critical foodproduction region of Indonesia. Awell defined system of regionalincentives and effective land usecontrols is to be incorporated withinthis economic developmentframework. These may include the

acquisition of development rightsand conservation easements to avertprime farmland conversion to non-agricultural uses or agricultural taxincentives to sustain a viable agricul-tural sector. Similarly, thegovernment may assist the agricul-tural sector in enhancing productionefficiencies and in the provision ofagricultural inputs (e.g., physicalinfrastructure, input subsidies, tech-nology transfer) and in the market-ing of agricultural products and byimproving the rural serviceinfrastructure. In addition, thecreation of value-added foodprocessing industries in rural areasmay assist in enhancing the econom-ic opportunities for those displacedfrom traditional agriculture,improve wage rates and reduce theeffects of rural push factors stimulat-ing urbanization. At the same time,this scenario will reduce agriculturalland conversion, which underminesregional agricultural sustainability.

48

References

Amin, Samir. 1974. Accumulation ona World Scale: A Critique of theTheory of Underdevelopment. 2vols. New York: Monthly ReviewPress.

Amin, Samir. 1976. UnequalDevelopment. New York:Monthly Review Press.

Andersen, Per Pinstrup, and PeterB.R. Hazell. 1985. The Impact ofthe Green Revolution andProspects for the Future. FoodReviews International, 1(1):1-25.

Armstrong, Warwick, and T.G.McGee. 1985. Theatres ofAccumulation: Studies in Asianand Latin American Urbanization.New York: Methuen & Co.

Atkinson, Adrian. 1992. The UrbanBioregion as SustainableDevelopment Paradigm. ThirdWorld Planning Review, 14(4):327-354.

Baran, Paul. 1979. On the PoliticalEconomy of Backwardness. InCharles K. Wilber (ed.), ThePolitical Economy ofDevelopment andUnderdevelopment. New York:Random House. Pp. 91-102.

Barber, Gerald M. 1988. ElementaryStatistics for Geographers. NewYork: The Guilford Press.

Batty, Michael. 1976. UrbanModelling: Algorithms,Calibrations, Predictions.Cambridge, England: CambridgeUniversity Press.

Beckford, George L. 1979.Comparative Rural Systems,Development andUnderdevelopment. In Charles K.Wilber (ed.), The PoliticalEconomy of Development andUnderdevelopment (second edi-tion). New York: Random House.

Berry, B.J.L., and A.M. Baker. 1968.Geographic Sampling. In B.J.LBerry and D.F. Marble, SpatialAnalysis: A Reader in StatisticalGeography. Englewood Cliffs,N.J.: Prentice-Hall. Pp. 24-34.

Berry, Brian J.L. 1964. Approaches toRegional Analysis: A Synthesis.Annals of the Association ofAmerican Geographers, 54:2-11.

Berry, Brian J.L. 1981. ComparativeUrbanization: Divergent Paths inthe Twentieth Century (revisededition). New York: St. Martin’sPress.

Bhadra, Dipasis, and A.S.P. Brandao.1993. Urbanization, AgriculturalDevelopment, and LandAllocation. World BankDiscussion Papers, No. 201.Washington, D.C.: The WorldBank.

Bhooshan, B.S., and R.P. Misra. 1980.Role of Small and IntermediateTowns in the Development ofMandya District. In B.S. Bhooshan(ed.), Toward AlternativeSettlement Strategies. New Delhi,India: Heritage Publishers.

Blair, John P. 1991. Urban andRegional Economics. Homewood,Ill., and Boston, Mass.: Irwin.

Booth, Anne (ed.). 1992. The OilBoom and After: IndonesianEconomic Policy and Performancein the Soeharto Era. New York:Oxford University Press.

Booth, Anne. 1988. AgriculturalDevelopment in Indonesia.Sydney, Australia: Allen andUnwin.

Breese, Gerald (ed.). 1966. City inNewly Developing Countries:Readings on Urbanism andUrbanization. Englewood Cliffs,N.J.: Prentice-Hall, Inc.

Brinser, Ayers. 1976. Essays onNatural Resource Management.Compiled and edited by Paul E.Nickel and Marylin M. Wallace.Winnipeg, Manitoba, Canada: TheNatural Resource Institute,University of Manitoba.

Bryant, C., and L.G. White. 1982.Managing Development in theThird World. Boulder, Colo.:Westview Press.

Choe, Sang-Chuel. 1981. Korea:From Chaos to Progress. Chapter5 in M. Honjo (ed.), Urbanizationand Regional Development.Maruzen Asia: United NationsCenter for Regional Development,Nagoya, Japan.

Chuta, Enyinna, and Carl Liedholm.1990. Rural Small-Scale Industry:Empirical Evidence and PolicyIssues. Chapter 19 in Carl K.Eicher and John M. Staatz (eds.),Agricultural Development in theThird World (second edition).Baltimore: Johns HopkinsUniversity Press. Pp. 327-341.

Coates, B.E., R.J. Johnson and P.L.Knox. 1977. Geography andInequality. New York: OxfordUniversity Press.

Colman, D., and F. Nixson. 1978.Economics of Change in LessDeveloped Countries. Oxford,England: Philip Allan.

Cowell, F. 1977. MeasuringInequality: Techniques for theSocial Sciences. New York: Wileyand Sons.

Cramer, Gail L., and Clarence W.Jensen. 1988. AgriculturalEconomics and Agribusiness(fourth edition). New York: JohnWiley and Sons.

Dadzie, K.K.S. 1980. EconomicDevelopment. Scientific American,Vol. 243, No. 3., September.

49

Datta, Satya. 1990. Why a NewApproach to Third WorldUrbanization. In Satya Datta (ed.),Third World Urbanization.Stockholm: Swedish Council forResearch in the Humanities andSocial Sciences. Pp. 9-23.

DESFIL. (No Date). Beyond theProject: The Quest forSustainability in the Third World.Working Paper of Developmentfor Fragile Lands. Washington,D.C.: USAID.

Dicken, Peter, and Peter E. Lloyd.1990. Location in Space (third edi-tion). New York: Harper CollinsPublishers.

Douglas, G.K. 1984. The Meaning ofAgricultural Sustainability. InG.K. Douglas (ed.), AgriculturalSustainability in a ChangingWorld Order. Westview SpecialStudies in Agriculture Science andPolicy. Pp. 3-29. Boulder, Colo.:Westview Press.

Dunford, Richard W. 1981. FarmlandRetention: Efficiency vs. Equity. InWalter E. Jeske (ed.), Economics,Ethics, Ecology: Roots ofProductive Conservation. Iowa:Soil Conservation Society ofAmerica. Pp. 367-379.

Eicher, Carl K., and John M. Staatz.1990. Agricultural Developmentin the Third World (secondedition). Baltimore, Md.: JohnsHopkins University Press.

El-Bushra, El-Sayed. 1989. TheUrban Crisis and Rural-UrbanMigration in Sudan. In R.B. Potterand T. Unwin (eds.), TheGeography of Urban-RuralInteraction in DevelopingCountries. London: Routledge.Pp. 109-140.

El-Shakhs, S. 1976. Planning forSystems of Settlements inEmerging Nations. Town PlanningReview, 7(1):127-138.

Ellis, G.F.R. 1980. Quality of LifeConcept: An Overall Frameworkfor Assessment Schemes. SaldruWorking Paper No. 30.

Enke, Stephen. 1962a.Industrialization through GreaterProductivity in Agriculture.Review of Economics and Statistics,44(February):88-91.

Enke, Stephen. 1962b. EconomicDevelopment with Limited andUnlimited Supplies of Labor.Oxford Economic Papers,14(June):158-172.

Erwidodo. 1990. Panel Data Analysison Farm-Level Efficiency, InputDemand and Output Supply ofRice Farming in West Java,Indonesia. A doctoral dissertation.Department of AgriculturalEconomics, Michigan StateUniversity, East Lansing, Mich.

Firman, Tommy. 1992. The SpatialPattern of Urban PopulationGrowth in Java, 1980-1990.Bulletin of Indonesian EconomicStudies, 28(2):95-109.

Frank, Andre Gunder. 1979. TheDevelopment ofUnderdevelopment. In Charles K.Wilber (ed.), The PoliticalEconomy of Development andUnderdevelopment. New York:Random House. Pp. 103-113.

Friedman, John. 1966. Two Conceptsof Urbanization. Urban AffairsQuarterly, 1:78-84.

Friedman, John. 1973. Urbanization,Planning, and NationalDevelopment. Beverly Hills,Calif.: Sage Publications.

Friedman, John (ed.). 1974. NewConcepts and Technologies inThird World Urbanization. LosAngeles, Calif.: University ofCalifornia Press.

Friedman, John. 1980. Urban Bias inRegional Development Policy. LosAngeles, Calif.: University ofCalifornia Press.

Friedman, John. 1988. Life Space andEconomic Space. New Brunswick,N.J.: Transaction Books.

Friedman, John, and Clyde Weaver.1979. Territory and Function: TheEvolution of Regional Planning.Berkeley and Los Angeles:University of California Press.

Friedman, John, and Goetz Wolff.1982. World City Formation: AnAgenda for Research and Action.Los Angeles, Calif.: University ofCalifornia Press.

Getis, Arthur, and Barry Boots. 1978.Models of Spatial Processes. NewYork: Cambridge UniversityPress.

Gillis M., D.H. Perkins, M. Roemerand D.R. Snodgrass. 1987.Economics of Development (sec-ond edition). New York: W.W.Norton & Company.

Gluck, Peter R., and R.C. Meister.1979. Cities in Transition: SocialChanges and InstitutionalResponses in UrbanDevelopment. New York: NewViewpoints, A Division ofFranklin Watts.

Goldberg, Michael, and PeterChinloy. 1984. Urban LandEconomics. Toronto: John Wiley &Sons.

Golstein, Sidney, and David Sly.1974. Basic Data Needed for theStudy of Urbanization.International Union for theScientific Study of Population.

Gottman, Jean. 1961. Megalopolis.Cambridge, Mass.: MassachusettsInstitute of Technology Press.

Gould, Peter R. 1972. PedagogicReview: Entropy in Urban andRegional Modelling. Annals of theAssociation of AmericanGeographers, 62:289-700.

50

Government of West Java(Pemerinath Propinsi DaerahTingkat I Jawa Barat). 1990. PolaDasar Pembangungan Daerah:Propinsi Daerah Tingkat I JawaBarat 1989/1990-1993/1994.Pemerintah Propinsi DaerahTingkat I Jawa Barat, Jakarta,Indonesia.

Hamer, A.M., A.D. Steer and D.G.Williams. 1986. Indonesia: TheChallenge of Urbanization. WorldBank Staff Working Papers, No.787. Washington, D.C.: The WorldBank.

Hardjono, Joan. 1983. RuralDevelopment in Indonesia: the‘top-down’ approach. Chapter 2In D.A.M. Lea and D.P. Chaudhri(eds.), Rural Development andthe State: Contradictions andDilemmas in DevelopingCountries. New York: Methuen &Co, Ltd.

Hardoy, J.E., and D. Satterthwaite.1986. Why Small andIntermediate Urban Centers? InJ.E. Hardoy and D. Satterthwaite(eds.), Small and IntermediateUrban Centers: Their Role inRegional and NationalDevelopment in the Third World.London: Hodder and Stoughtonin association with theInternational Institute forEnvironment and Development.Pp. 1-17.

Harrison, Lawrence E. 1985.Underdevelopment is a State ofMind. The Center forInternational Affairs, HarvardUniversity. Madison Books.

Hauser, Philip M. 1985. Introductionand Overview. Chapter 1 in PhilipM. Hauser, Daniel B. Suits andNaohiro Ogawa (eds.),Urbanization and Migration inAsian Development. Tokyo:National Institute for ResearchAdvancement.

Hauser, Philip M., Daniel B. Suitsand Naohiro Ogawa. 1985.Urbanization and Migration inAsian Development. Tokyo:National Institute for ResearchAdvancement.

Hayami, Y. and V.W. Ruttan. 1985.Agricultural Development: AnInternational Perspective (revisedand expanded edition). Baltimore,Md.: The Johns HopkinsUniversity Press.

Hehanussa, P.E., and F. Hehuwat.1979. Morphogenesis of theNorthern Coastal Plain of WestJava between Cirebon andJakarta: Its Implications forCoastal Zone Management. InE.C.F. Bird and A. Soegiarto (eds.),Proceedings of the JakartaWorkshop on Coastal ResourcesManagement. Pp. 22-30.

Helmer, John, and Neil A. Eddington(eds.). 1973. Urban man: ThePsychology of Urban Survival.New York: Free Press, a divisionof Macmillan Publishing Co., Inc.

Henderson, J.V. 1977. EconomicTheories and the Cities. NewYork: Academic Press.

Hill, Hal. 1992a. Regional develop-ment in a Boom and BustPetroleum Economy: Indonesiasince 1970. Economic Developmentand Cultural Change, 40(2):351-377.

Hill, Hal. 1992b. ManufacturingIndustry. Chapter 7 in Anne Booth(ed.), The Oil Boom and After:Indonesian Economic Policy andPerformance in the Soeharto Era.New York: Oxford UniversityPress.

Hirschman, Albert O. 1958. TheStrategy of EconomicDevelopment. New Haven,Conn.: Yale University Press.

Honjo, M. 1980. Role of Governmentin Regional Development. RegionalDevelopment Dialogue, 1:1-16.

Honjo, M. 1981a. Introduction.Chapter 1 in M. Honjo (ed.),Urbanization and RegionalDevelopment. Maruzen Asia:United Nations Center forRegional Development, Nagoya,Japan.

Honjo, M. 1981b. Overview ofUrbanization andMetropolitanization in Asia.Chapter 2 in M. Honjo (ed.),Urbanization and RegionalDevelopment. Maruzen Asia:United Nations Center forRegional Development, Nagoya,Japan.

Hugo, Graeme J. 1978. PopulationMobility in West Java. Yogyakarta:Gadjah Mada University Press.

Ikerd, J.E. 1989. Conflicts betweenSustainability and Profitability inAgriculture. Unpublished paperin seminar series in sustainabledevelopment. Michigan StateUniversity. Winter 1990.

Ilahude, A.G. 1979. TheOceanographic Features of theCoastal Region between Jakartaand Cirebon: Oceanography. InE.C.F. Bird and A. Soegiarto (eds.),Proceedings of the JakartaWorkshop on Coastal ResourcesManagement. Pp. 31-35.

James, William E. 1985. ASEANAgriculture and EconomicDevelopment. Chapter 2 in P.M.Hauser (ed.), Urbanization andMigration in Asian Development.Tokyo: National Institute forResearch Advancement.

Johnston, B.F., and J.W. Mellor. 1961.The Role of Agriculture inEconomic Development. AmericanEconomic Review, 51:566-593.

Johnston, R.J. 1991. MultivariateStatistical Analysis in Geography.New York: John Wiley and Sons,Inc.

51

Jorgenson, Dale W. 1961. TheDevelopment of a Dual Economy.Economic Journal, 72(June):309-334.

Jorgenson, Dale W. 1969. The Role ofAgriculture in EconomicDevelopment: Classical versusNeoclassical Models of Growth.In Clifton R. Wharton Jr.,Subsistence Agriculture andEconomic Development. Chicago:Aldine. Pp. 320-48.

Kabwegyere, T.B. 1979. Small UrbanCenters and the Growth ofUnderdevelopment in RuralKenya. In A. Southall (ed.), SmallUrban Centers in RuralDevelopment in Africa. Madison,Wis.: African Studies Program,University of Wisconsin. Pp. 36-44.

Kelley, Allen C., and Jeffrey G.Williamson. 1982. The Limits ofUrban Growth: Suggestions forMacro Modelling Third WorldEconomies. Economic Developmentand Culture Change, 30(April):595-623.

Killen, James. 1983. MathematicalProgramming Methods forGeographers and Planners. NewYork: St Martin’s Press.

Kleinbaum, David G., and LawrenceL. Kupper. 1978. AppliedRegression Analysis and OtherMultivariable Methods. NorthScituate, Mass.: Duxbury Press.

Knox, Paul L. 1974. Level of Living:A Conceptual Framework forMonitoring Regional Variations inWell-being. Regional Studies, 8:11-19.

Kuz, Tony J. 1978. Quality of Life: AnObjective and Subjective VariableAnalysis. Regional Studies, 12:409-417.

Land, Kenneth C., and SeymourSpilerman (eds.). 1975. SocialIndicator Model. New York:Russell Sage Foundation.

Lele, S.M. 1991. SustainableDevelopment: A Critical Review.World Development, 19(6):607-621.

Lipton, M. 1982. Why Poor PeopleStay Poor. In J. Harris (ed.)., RuralDevelopment: Theories of PeasantEconomy and Agrarian Change.London: Hutchinson. Pp. 66-81.

Little, Ian M.D. 1982. EconomicDevelopment: Theory, Policy, andInternational Relations. NewYork: Basic Books, Inc.

Lockeretz, William. 1986. AlternativeAgriculture. Chapter 13 inKenneth A. Dahlberg (ed.), NewDirections for Agriculture andAgricultural Research: NeglectedDimensions and EmergingAlternatives. New Jersey:Rowman & Allanheld Publishers.Pp. 291-311.

Mabogunje, A.L. 1981. TheDevelopment Process: A SpatialPerspective. New York: Holmesand Meier Publications.

Manning, Chris. 1988. The GreenRevolution, Employment, andEconomic Change in Rural Java:Reassessment of Trends under theNew Order. Asian EconomicResearch Unit, Institute ofSoutheast Asian Studies.

Marshall, John U. 1989. TheStructure of Urban Systems.Toronto: University of TorontoPress.

Matthews, Ralph. 1983. The Creationof Regional Dependency. Toronto:University of Toronto Press.

McAndrew, John. 1990. The UrbanProcess and Social Differentiationin a Philippine Village. In SatyaDatta (ed.), Third WorldUrbanization. Stockholm:Swedish Council for Research inthe Humanities and SocialSciences. Pp. 186-203.

McGee, T.G. 1967. The SoutheastAsian Cities. London: G. Bells andSon.

McGee, T.G. 1971. The UrbanizationProcess in the Third World:Explorations in Search of aTheory. London: G. Bell and Sons,Ltd.

McGee, T.G. 1989. Urbanisasi orKotadesasi? Evolving Patterns ofUrbanization in Asia. Chapter 6 inF.J. Costa, A.K. Dutt, L.J.C. Maand A.G. Noble (eds.),Urbanization in Asia: SpatialDimensions and Policy Issues.Honolulu: University of HawaiiPress. Pp. 93-108.

McGee, T.G. 1991. The Emergence ofDesakota Regions in Asia:Expanding a Hypothesis. In N.Ginsburg, B. Koppel and T.G.McGee (eds.), Metropolis:Settlement Transition in Asia.Honolulu: University of HawaiiPress. Pp. 3-26.

Mehretu, Assefa. 1989. RegionalDisparities in Sub-Saharan Africa:Structural Readjustment ofUneven Development. Boulder,Colo.: Westview Press.

Mehretu, Assefa, R.I. Wittick andB.W. Pigozzi. 1983. Spatial Designfor Basic Needs in Eastern UpperVolta. Journal of Developing Areas,17:383-394.

Mehretu, Assefa. 1983. Cities ofSubsaharan Africa. In Stanley D.Brunn and Jack F. Williams (eds.),Cities of the World: WorldRegional Urban Development.New York: Harper CollinPublishers, Inc. Pp. 243-280.

Mellert, R. 1985. Why We ShouldFeel Responsible for FutureGenerations. In D.D. Chiras,Environmental Science: AFramework for Decision Making.The Benjamins/CummingsPublishing Co., Inc.

Mellor, John W. 1966. The Economicsof Agricultural Development.Ithaca, N.Y.: Cornell UniversityPress.

52

Mellor, John W. 1986. Agriculture onthe Road of Industrialization. InJohn P. Lewis and ValerianaKallab (eds.), DevelopmentStrategies Reconsidered. OverseasDevelopment Council: U.S.-ThirdWorld Policy Perspectives No. 5.New Brunswick: TransactionBooks. Pp. 67-90.

Mellor, John W. 1988. Global FoodBalances and Food Security. WorldDevelopment, 16(9):997-1011.

Moochtar, Radinal. 1992. Keynoteaddress: “Urban Growth towardsthe Year 2000 and Beyond.”Ministry of Public Works, Jakarta,19 November 1992.

Mullen, John D., and Byron M. Roth.1991. Decision Making: Its Logicand Practice. Maryland: Rowman& Littlefield Publishers, Inc.

Murphey, Rhoads. 1966.Urbanization in Asia. Chapter 6 inG. Breese (ed.), The City in NewDeveloping Countries: Readingson Urbanism and Urbanization.Englewood Cliffs, N.J.: Prentice-Hall, Inc. Pp. 58-75.

Myrdal, Gunnar. 1957. Rich Landsand Poor: The Road to WorldProsperity. New York: Harper.

Nasoetion, L.I. 1994. KebijaksanaanPertanahan Nasional dalamMendukung PembangunanEkonomi: Pengalaman Masa Lalu,Tantangan dan Arah ke MasaDepan. Orasi Ilmiah Guru BesarTetap Ilmu Tanah, FakultasPertanian, Institut PertanianBogor, Bogor, Indonesia.

Nath, V. Urbanization in India:Retrospect and Prospect. Chapter3 in F.J. Costa, A.K. Dutt, L.J.C.Ma and A.G. Noble (eds.),Urbanization in Asia: SpatialDimensions and Policy Issues.Honolulu: University of HawaiiPress. Pp. 27-46.

Naylor, Rosamond. 1992. Labor-Saving Technologies in theJavanese Rice Economy: RecentDevelopments and a Look intothe 1990s. Bulletin of IndonesianEconomic Studies, December 1992,28(3):71-92.

Neter, John, and WilliamWasserman. 1974. Applied LinearStatistical Models: Regression,Analysis of Variance andExperimental Designs.Homewood, Ill.: Richard D. Irwin,Inc.

Obudho, Robert A., and Waller,Peter P. 1976. Periodic Markets,Urbanization and RegionalPlanning. Westport, Conn.:Greenwood Press.

Okun, B., and R.W. Richardson(eds.). 1961. Studies in EconomicDevelopment. New York: Holt,Rinehart and Winston.

Ongkosongo, Otto S.R. 1979. TheOceanographic Features of theCoastal Region between Jakartaand Cirebon: Geology. In E.C.F.Bird and A. Soegiarto (eds.),Proceedings of the JakartaWorkshop on Coastal ResourcesManagement. Pp. 35-42.

Ott, Lyman. 1988. An Introduction toStatistical Methods and DataAnalysis (third edition). Boston:PWS-Kent Publishing Company.

Phillips, Paul. 1978. RegionalDisparities. Toronto: JamesLorimer & Co.

Pingali, P.L., P.F. Moya and L.E.Velasco. 1990. The Post-GreenRevolution Blues in Asian RiceProduction: the Diminished GapBetween Experiment Station andFarmer Yields. Social ScienceDivision, International RiceResearch Institute (IRRI), Manila,Philippines.

Poot, H., A. Kuyvenhoven and J.Jansen. 1990. Industrializationand Trade in Indonesia.Yogyakarta: Gadjah MadaUniversity Press.

Preston, D. 1975. Rural-Urban andInter-Settlement Interaction:Theory and Analytical Structure.Area, 7:171-174.

Ranis, Gustav, and John C.H. Fei.1961. A Theory of EconomicDevelopment. American EconomicReview, 51(4):533-565.

Ranis, Gustav. 1963. Innovation,Capital Accumulation andEconomic Development. AmericanEconomic Review, 53(3):283-313.

Ranis, Gustav. 1964. Development ofthe Labor Surplus Economy:Theory and Policy. Homewood,Ill.: Richard D. Irwin.

Reed, Robert R. 1990. Patterns ofMigration in Southeast Asia.Center for South and SoutheastStudies, University of Californiaat Berkeley, California.

Reissman, Leonard. 1964. The UrbanProcess: Cities in IndustrialSocieties. Illinois: Glencoe.

Rondinelli, D.A. 1983. SecondaryCities in Developing Countries:Policies for DiffusingUrbanization. Beverly Hills,Calif.: Sage.

Rondinelli, D.A. 1985. AppliedMethods of Regional Analysis: theSpatial Dimensions ofDevelopment Policy. Boulder,Colo.: Westview.

Rondinelli, Denis A., and KennethRuddle. 1976. Urban Functions inRural Development: An Analysisof Integrated SpatialDevelopment Policy. Washington,D.C.: Office of UrbanDevelopment, TechnicalAssistance Bureau, Agency forInternational Development, U.S.Department of State.

53

Runyon, Richard P., and AudreyHaber. 1980. Fundamentals ofBehavioral Statistics.Massachusetts and California:Addison-Wesley PublishingCompany.

Ruttan, Vernon W. 1990. Models ofAgricultural Development. InCarl K. Eicher and John M. Staatz(eds.), Agricultural Developmentin the Third World (secondedition). Baltimore, Md.: JohnsHopkins University Press. Pp. 89-96.

Ruttan, Vernon W. 1990. SustainableIs Not Enough. Chapter 24 in CarlK. Eicher and John M. Staatz(eds.), Agricultural Developmentin the Third World (secondedition). Baltimore, Md.: JohnsHopkins University Press. Pp.400-404.

Santos, T.D. 1970. The Structure ofDependence. American EconomicReview, 60:231-236.

Schaik, A.V. 1986. Colonial Controland Peasant Resources in Java:Agricultural InvolutionReconsidered. NetherlandsGeographical Studies(Nederlandse GeografischeStudies) 14.

Schickele, Rainer. 1954. AgriculturalPolicy: Farm Programs andNational Welfare. New York:McGraw-Hill Book Company, Inc.

Schmid, A. Alan. 1968. ConvertingLand from Rural to Urban Uses.Baltimore, Md.: Johns HopkinsPress.

Schultink, Gerhardus. 1992.Evaluation of SustainableDevelopment Alternatives:Relevant Concepts, ResourceAssessment Approaches andComparative Spatial Indicators.Intern. J. Environmental Studies,41:203-234.

Seers, Dudley. 1970. The Meaning ofDevelopment. ADC Reprint. NewYork: Agricultural DevelopmentCouncil.

Sen, Amartya. 1990. Food,Economics, and Entitlements.Chapter 11 in Carl K. Eicher andJohn M. Staatz (eds.), AgriculturalDevelopment in the Third World(second edition). Baltimore, Md.:Johns Hopkins University Press.Pp. 189-205.

Sharbatoglie, A. 1991. Urbanizationand Regional Disparities in Post-Revolutionary Iran. Boulder,Colo., San Francisco, Calif., andOxford: Westview Press.

Sjoberg, Gideon. 1960. The Pre-industrial City: Past and Present.New York: The Free Press.

Smith, David M. 1982. Where theGrass is Greener: Living in anUnequal World. Baltimore, Md.:Johns Hopkins University Press.

Soegijoko, Sugijanto. 1992. Regionaland Urban Development inIndonesia: Achievement andOptions for the Future. Keynotepaper in the conference “UrbanGrowth towards the Year 2000and Beyond.” Jakarta, 19November 1992. Pp. 7-28.

Soemarwoto, O., I. Soemarwoto, H.Salim, O.S. Abdoellah, C. Asdak,R. Muslihudin, S. Astuti and T.Herawati. 1991. SustainableDevelopment of West Java.Institute of Ecology, PadjadjaranUniversity, Bandung, Indonesia.

Soemarwoto, Otto. 1992. Questionsof Sustainability in UrbanDevelopment. Paper presented inthe conference “Urban Growthtowards the Year 2000 andBeyond.” Jakarta, 19 November1992. Pp. I.28-I.44.

Southall, A.S. 1977. Small UrbanCenters in Rural Development inAfrica. Research proposal,University of Wisconsin, Madison(mimeographed).

Southall, Aidan. 1979. Introduction.In Aidan Southall, Small UrbanCenters in Rural Development inAfrica. African Studies Program,University of Wisconsin,Madison, Wis. Pp. 1-20.

Sovani, N.V. 1964. The Analysis ofOver-Urbanization. EconomicDevelopment and Cultural Change,12(12):113-122.

Stevens, Robert D., and Cathy L.Jabara. 1988. AgriculturalDevelopment Principles:Economic Theory and EmpiricalEvidence. Baltimore, Md.: JohnsHopkins University Press.

Stohr, Walter B. 1981. Developmentfrom Below: The Bottom-Up andPeriphery-Inward DevelopmentParadigm. Chapter 2 in Walter B.Stohr and D.R.F. Taylor,Development from Above orBelow? The Dialectics of RegionalPlanning in DevelopingCountries. New York: John Wileyand Sons. Pp. 39-72.

Streeten, Paul. 1984. Basic Needs:Some Unsettled Questions.Working paper No. 56, MichiganState University.

Sutton, Keith. 1989. The Role ofUrban Bias in Perpetuating Rural-Urban Disparities in the Maghreb.In R.B. Potter and T. Unwin (eds.),The Geography of Urban-RuralInteraction in DevelopingCountries. London: Routledge.Pp. 68-108.

Tabor, Steven R. 1992. Agriculture inTransition. Chapter 6 in AnneBooth (ed.), The Oil Boom andAfter: Indonesian EconomicPolicy and Performance in theSoeharto Era. New York: OxfordUniversity Press. Pp. 161-203.

Tarrant, J. et al. 1987. NaturalResources and EnvironmentalManagement in Indonesia.Jakarta, Indonesia: United StatesAgency for InternationalDevelopment.

54

Thomas, R.W., and R.J. Hugget.1980. Modelling in Geography: AMathematical Approach. Totowa,N.J.: Barnes & Noble Books.

Thrall, Grant Ian. 1987. Land Useand Urban Form: TheConsumption Theory of LandRent. New York: Methuen.

Timmer, C. Peter. 1990. TheAgricultural Transformation.Chapter 2 in Carl K. Eicher andJohn M. Staatz (eds.), AgriculturalDevelopment in the Third World(second edition). Baltimore, Md.:Johns Hopkins University Press.Pp. 47-69.

Toner, William. 1979. Local Programsto Save Farms and Farmlands. InMax Schnepf (ed.), Farmland,Food, and the Future. Iowa: SoilConservation Society of America.Pp. 189-202.

UNCED. 1987. Our Common Future.Oxford University Press.

Unwin, Tim. 1989. Urban-RuralInteraction in DevelopingCountries: A TheoreticalPerspective. Chapter 1 in R.B.Potter and Tim Unwin (eds.), TheGeography of Urban-RuralInteraction in DevelopingCountries. London: Routledge.Pp. 11-32.

Wilber, Charles K., and James H.Weaver, 1979. Patterns ofDependency: Income Distributionand the History ofUnderdevelopment. In Charles K.Wilber (ed.), The PoliticalEconomy of Development andUnderdevelopment. New York:Random House. Pp. 114-130.

Williams, Frederick. 1986. Reasoningwith Statistics (third edition).New York: Holt, Rinehart andWinston.

Williams, Jack F., Stanley D. Brunnand Joe T. Darden. 1983. WorldUrban Development. In StanleyD. Brunn and Jack F. Williams(eds.), Cities of the World: WorldRegional Urban Development.New York: Harper CollinPublishers, Inc. Pp. 3-42.

Winoto, Joyo. 1985. “PengaruhPermukaan Air Tanah TerhadapProses Pembentukan TanahSawah” (Impacts of Water Tableson the Genesis of Paddylands).Department of Soil Sciences,Bogor Agricultural University,Bogor, Indonesia.

World Bank. 1990. Indonesia:Strategy for a SustainedReduction in Poverty. A WorldBank Country Study. Washington,D.C.: The World Bank.

World Bank. 1994. WorldDevelopment Report:Infrastructure for Development.The International Bank forReconstruction andDevelopment/The World Bank.Washington, D.C.: The WorldBank.

World Handbook, 1995. Washington,D.C.: U.S. Central IntelligenceService.

Wrigley, N. (ed.). 1979. StatisticalApplications in the SpatialSciences. Brondesbury Park,London: Page Bros. (Norwich),Limited.

Zain, Z.B.M. 1987. Pattern of SpatialDevelopment and RegionalInequalities in PeninsularMalaysia. A dissertation. Urbanand Regional Planning Program,College of Social Science,Michigan State University.

55

The Michigan Agricultural Experiment Station is an equal opportunity employer and complies with Title VI of the Civil Rights Act of1964 and Title IX of the Education Amendments of 1972. New 4:96 - TCM - 1M