an analytical survey of population and …an analytical survey ofpopulation and development in...

An Analytical Survey of Population and Development in Bangladesh

Arthur, W.B. and McNicoll, G.

IIASA Professional PaperSeptember 1978

Arthur, W.B. and McNicoll, G. (1978) An Analytical Survey of Population and Development in Bangladesh. IIASA

Professional Paper. IIASA, Laxenburg, Austria, PP-78-008 Copyright © September 1978 by the author(s).

http://pure.iiasa.ac.at/916/ All rights reserved. Permission to make digital or hard copies of all or part of this

work for personal or classroom use is granted without fee provided that copies are not made or distributed for

profit or commercial advantage. All copies must bear this notice and the full citation on the first page. For other

purposes, to republish, to post on servers or to redistribute to lists, permission must be sought by contacting

[email protected]

mailto:[email protected]

PP-78-8

AN ANALYTICAL SURVEY OF POPULATIONAND DEVELOPMENT IN BANGLADESH

w. Brian ArthurGeoffrey McNicoll

September 1978

ProfessionalPapersare not official publicationsof the InternationalInstitutefor Applied Systems Analysis, but are reproducedand distributed by theInstitute as an aid to staff membersin furthering their professionalactivities.Views or opinionsexpressedhereinare thoseof the authorand shouldnot beinterpretedas representingthe view of either the Institute or the NationalMemberOrganizationssupportingthe Institute.

An Analytical Surveyof Populationand Developmentin Bangladesh

W. BRIAN ARTHURGEOFFREYMeN/COll

It is helpful at times to standbackfrom a subjectand try to pull togetherthe various strandsof knowledgethat haveaccumulatedabout it. This is our purposein the presentstudyof populationand developmentin Bangladesh.In writing this surveywepresentmore than a collection of facts and figures. Throughout,we areconcernedwith analysisof the forces and realitiesbehindthe data,withtrying to makesenseof the economicand demographicprocessesthat liebeneathBangladesh'sdevelopmentproblems.

We bring to the studyno new evidenceor data; rather,we marshalexisting knowledgeas a basefor analyzingpopulationand developmentissuesand assessingdevelopmentalternatives.Becauseof the sometimesweak and fragmentarynatureof this knowledgein the caseof Bangla-desh, the picture that emergesis in placesblurred and speculative.Butit is consistentenoughon the whole to allow us to draw somebroadcon-clusions.A country study might be expectedto forecasttrendsand offerprescriptionsfor policy. Herewe do neither.Our interestis moreto locatethe particulareconomic,demographic,and social forces that seemlikelyto influence the style of future developmentand set its bounds.Ratherthan proposespecific measuresin a country where governmenthas, atbest,a limited impacton local realities,we keeppolicy discussionat thelevel of overall strategy.

Economicand demographicbehavioris usually seenin eithermacroor micro terms. National statisticsare the outcomeof decisionsat theindividual level, and individual incentivesin tum are set by aggregate

23

24 POPULATION AND DEVELOPMENT IN BANGLADESH

conditions.Particularly in Bangladesh,however,we must seekmuch ofour understandingat a level betweentheseextremes-inthe details ofsocial organizationand local institutions. We spendconsiderabletimediscussingthe Bangladeshsituation at this intermediatelevel. It is herethat problemsfacedby governmentprograms,prospectsfor rural change,and present adverse patterns of demographicbehavior can best beexplained.

We beginwith a view of Bangladeshat the aggregatelevel, exanlin-ing where the country standsand how it cameto that position. In thesecondandthird sectionswe shift our focusto the local level, investigatinghow socialorganizationaffectsand is affectedby the local economy,andhow thesein tum influence individual incentivesand demographicbe-havior. A final section assessesprospectsand possibilities for futuredevelopmentin Bangladesh.

The BangladeshSetting

The economiccircumstancesof Bangladeshare as familiar as they areprecarious.Per capita incomesare amongthe lowest in the world; lessthan 15 percentof the adult population has had five or more years'schooling;farm holdingsare small and fragmented,with a growing pro-portion (now perhapsa third) of rural householdslandless;the countryhas not been self-sufficient in food for over two decades;nutritionallevelsarebelievedto beinadequatefor half the population.Demographicconditions,too, are extreme.Crowdedinto a country a third the size ofCalifornia aresome85 million people;over 75 million of them live in thecountryside,making Bangladeshthe most denselypopulatedrural coun-try in the world. Marriagetakesplaceat a very young ageand is all butuniversal;total fertility, the averagenumberof childrenborn to a womansurviving throughher reproductiveyears,may be as much as seven;anddeathratesarestill high, though perhapshalf the level of 25 years ago.Evenoptimistically, the populationgrowth ratecannotbe lower than 2.5percentper year, a level that would double the population in under30years.

What is interestingaboutthis litany of grim statisticsis not that lifein Bangladeshis at a baresubsistencelevel but that economicanddemo-graphicbehaviorworseningtheseconditionscan persist in the face ofsuchcircumstancesandmay indeedbe fosteredby them. In what followswe investigatehow and why this is so.

Colonial Past

For mostof the last four centuries,as Bertocci (1976) hasremarked,thesetting for Bangladesh'seconomichistory has been"the control of East

W. Brian Arthur / GeoffreyMcNicoll 25

"

Bengalby structuresof statepower whosedecision-makingcentershavebeenoutsidethe region." For some200 yearsfollowing Bengal'sformalincorporationin the Moghul empirein 1576, the region seeminglymain-tained a relative prosperitythat, to judge from travelers'reports,datedback to the middle ages.l Rice cultivation dominatedthen as it doestoday,but a diversified agricultureand a successfullocal silk and cottonindustry were the basisof a vigorous export trade.

In the late eighteenthcentury,Moghul power was graduallyyieldedto the British East India Company, with formal British administra-tion following in 1857. Company dominance coincided with a rapiddeclinein local industry,tracingthe familiar patternof colonial territoriesin the Industrial Revolution but here hastenedby a series of naturaldisasters.From being a balancedagrarianeconomythe region was con-vertedinto a raw-materialsbasefor British industry. Local textile manu-facturing virtually ceasedin competitionwith machine-madeyams andfabrics,while specializationin export cropswas emphasized-indigoandsafBowerin the early1800s,later jute andtea.:2Underthis deindustrializa-tion, the populationof the main commercialcentersdeclined.3

Creationof Pakistanin 1947,accompaniedby the violenceand mas-sive migrationsof Partition, put an end to official colonial rule. But forEast Bengal (now East Pakistan),the new arrangementsmerely trans-ferred power from Delhi to Karachi, creating in effect a new colonialrelationshipthat did little to aid developmentin the region. Jutemilling,previouslyconcentratedaroundCalcuttaandcut off in the Partition,wasreestablishedin Dacca,but little elsehappenedto fulfill theearly promiseof economicprogress.Pakistan'sdevelopmentprogramwas ill suited tothe agrarianneedsof its EastWing. Agriculture stagnatedin the 195Os,picking up only slowly in the 1960swith the first stirrings of the GreenRevolution.Outputfailed to stayabreastof populationgrowth: Domesticdemandfor rice beganto shrink the area available for jute cultivation,and even so, by the late 1960s East Pakistanwas importing about 10percentif its food requirements.By this time, when moveswere finallymadeto remedysomeof the East-Westinequities,it was too late to avertcivil war.

In 1971 the new nation of Bangladeshemerged,following a nine-monthstrugglefor independencethat took asmanyashalf a million lives,displacedat leasttemporarilya tenth of the population,and saw the lossof a significant part of the country's physical capital and managerialtalent:'

One continuing reality lay behind thesechangingregimesover thetwo centuriesprecedingindependence:Much of the region'sagriculturalsurplus was removed, leaving little available for investmentwithin it.Under the Companyand the fonnal British administration,this was thedeliberate outcome of colonial policy, rationalized in the values ofempire.Bengalhada significantgrain surplus,which the British managed


to extractby tighteningthe looseprocurementsystemthey inheritedfromthe Moghuls.UnderPakistanin the 1950sa shrunkenagriculturalsurpluscould still be removed,but now the mechanismswere different. Invest-ment funds wereallocatedpredominantlyto promoteindustrial develop-ment in Pakistan'sWest Wing, and the overvaluedcurrency,maintainedso asto favor capital-goodsimports,penalizedthe jute-exporteconomyinthe East. This was largely the result of a strategyof industrializationthroughimport substitution,popularin the 1950s,thatpromotedindustryat the expenseof agriculture. But in Pakistan the sectoral imbalancetypically createdby such a strategycoincided with the country's East-Westdivision. Much of the valueof the East'sjute exportswas in effecttransferredto the West.5

With the steadyremoval of potential invesbnentresourcesunder asuccessionof colonial regimes,the region had little chanceto developitseconomicbase.Indeed,growthof the rural economyin the yearsfrom theestablishmentof jute cultivation in the mid-nineteenthcentury throughthe 1960scould well be describedby the term that the Dutch economistJ. H. Boekeusedof Java:staticexpansion.Theeconomysteadilyenlargeditself, but its underlyingstructureremainedthe same.Bangladeshtodayretainsessentiallythe sameeconomicbasethat had beenestablishedbythe 185Os:cultivation of rice, jute, and a few othercrops, and export ofagricultural goods, mainly raw jute and manufacturedjute products.Even though total output is larger, so too is population,with the resultthattheimportantqualitativeindexes-cropyields, laborproductivity,andrealincomes-haveshownslight improvementor evenhaveworsenedovertheyears.6

If Bangladeshwasleft little in the way of useful economicstructurefrom its colonial era,neitherdid it inherit a strongadministrativeframe-work. Underthe "PermanentSettlement"of 1793 the British createda defacto landlord class, the zamindars,chargedwith collecting revenue.TThesein tum broughtinto being a multilayeredhierarchyof tenantsandundertenantsto squeezerevenueupward from the cultivator at the bot-tom. Authority at the village level came to rest in the hands of thesemiddle men, supplantingan alreadyweak systemof village self-govern-ment and acting in lieu of a formal administrativestructure. It was asystem designedto extract surplus, which it did easily, though withconsiderableleakage;but it left the British with minimal control or eveninformationbelow the level of the zamindars.Governmentreportsin thenineteenthcenturyspeakof "administrativestarvation" in East Bengal,a truncationof lines of administrationevenby the usual British colenialpatternof indirect rule. This condition persisteddespitea successionofefforts to fill the institutional gap.8 In the first decadeof Pakistanrulelocal governmentwas again neglected.Only with the introduction in1959 of the "Basic Democracies"system,which set up electedand ap-


pointed councils at various administrativelevels, did the outlines of afOlmal local governmentstmcturebegin to take shape.But the realitieswere disappointing. The national government'scapacity to influenceeventsat the local levelremainslimited today.

In sum, then, Bangladeshinherited little from its colonial era thatcould.be usedas a basisfor development.Two centuriesof underinvest-ment left an economy narrowly based, rural, and static. It also left alegacyof underadministration,so that when there was finally a commit-ment to rapid development,the national governmenthad limited capa-bility to promote neededchangesat the village level. The main challengeof creatingan appropriateframeworkfor nationaldevelopmenthasstill tobemet.

EnvironmentalConditions

Crucial to understandingmuch of what happensin Bangladeshare itsunique environmentalcircumstances.!lThe country lies on the delta ofthe Ganges,Brahmaputra,and Meghna rivers, an immense system oftributariesand distributariesdraining the central and easternHimalayasand the foothills of Assam. With the exceptionof the ChittagongHillTractsadjoining the borderswith India andBurma in the southeastandsome small areasin the northeast,most of the country is less than 15metersabovesealevel. In normal yearsa third of the cultivated area isonemeteror moreunderwaterat thepeakof the monsoonfloods. The siltdepositedduring this seasonalinundationmaintainsthe high soil fertility,making possiblethe extremerural-populationdensitiesthat characterizethe active delta areas(seeFigure 1).

Siltation has another, less favorable effect, however. It steadilyraisesthe bedsof therivers andthusrenderstheir coursesunstable.Thereare both short-run and long-mn consequences.In each monsoonsomebanksare eatenaway, while new land is built up elsewhere-sometimesonly to disappearagainin a future season.Quite extensiveland areasmaybe involved in theseyear-to-yearshifts, with a large impact on the ｬｯ｣｡ｾ

economy.Over the centuriesthere has been a steadyeastwardshift inthe active delta region, leaving major parts of the westerndistricts (theso-calledmoribund delta area) no longer flushed and renewedby theannual flooding. This trend has been marked by spectacularshifts insome of the main channels-mostrecently in the late 1700s and early1800s,when the Brahmaputrachangedcourseto form the JamunaRiverand the junction betweenthe Ganges(Padma)and Meghnarivers wasfirst formed.

With a very densepopulationliving undermarginalgeographiccon-ditions, it takesonly slight changesto producedisasters.Shifting river

28

Figure 1Bangladesh

POPULATION AND DEVELOPMENT IN BANGLADESH

OF

BAY OF BENGAL

o ｾｉｯｯ､･､ area

IT] l'iill country

o SO 100 Miloo1-1--...., __I _,-_...'

o SO 100 ISO KiIomocon

W. Brian Arthur / GeoffreyMcNicoU 29

coursesareonly oneof the naturalhazardsfacing Bangladesh.Variationsin the timing and amountof the monsoonrains can causeseverecropdamageand consequentlocal famine. With multiple cropping and thecontractedgrowing seasonsof some of the new crop varieties, thesefluctuations becomeevenmore ｾｲｩｴｩ｣｡ｬＮ Along the southerncoast in thetidal areas,salinity posesa constantthreat to agriculture. Cyclonesarecommonin the Bay of Bengal and haveaccountedfor much loss of lifeover the yearsthrough excessiveflooding. The tidal waves (bores) thatsometimesaccompanythem havebeen a major causeof destructiononthe islands in the delta mouth. Table 1 cpmts the worst of the naturaldisastersover the last two centuries,togetherwith theman-madecalami-tiesof morerecentyears.

Table1Major Disastenin Bangladeshin the last200Yean

1784-88 Floods and famine;radical shift in courseof Brahmaputra (1787)

1873-74 Famine

1876 Bakarganj Cyclone andtidal wave

1884-85 Famine

1897 Chittagong Cyclone

1918-19 Influenza epidemic

1943 Bengal Famine

1947 Partition of India

1970 Cyclone and tidal wave

1971 VVarofindependence

1974 Famine

Year

1769-76

Event

Great Bengal Famine

Casualties

"Eliminated almost a third of Bengal'spopulation" (A. Ahmed, 1962, p. 140),although impact was less severe in EastBengal (N. Ahmad, 1968, p. 327)

Unknown (N. Ahmad, 1968, pp. 33, 101)

Unknown (A. Ahmed, 1962, p. 141)

c. 400,000 deaths (N. Ahmad, 1968, p. 51)

Unknown (Bhatia, 1967, p. 164)

c. 175,000 deaths (N. Ahmad, 1968, p.51)

c. 400,000 deaths (M. R. Khan, 1972b,p.384)

2-2.5 million deaths (A. Ahmed, 1962,p. 141; M. R. Khan, 1972b, p. 384)

Unknown; total deaths in Partitionc. one million, but most were in VVest(Davis, 1951, p. 197)

200,000-500,000 deaths (L. C. Chen, 1973)

c. 500,000 deaths (Curlin, Chen, andHussain, 1976, p. 31)

Officially c. 30,000 deaths (Majlis, 1977),although some estimates are much higher(e.g., 500,000 - Baldwin, 1977; 80,000 inRangpur district alone - Haque et aI.,1977)


Environmentalcircumstances-inparticular, duration and depth ofHooding-largely dictate agricultural possibilities. They govern wheredoublecroppingis feasible,wherepaddycanbe transplantedratherthanbroadcast,and where short-stemmedvarieties rather than deep-waterHoating rice can be grown-all factors that greatly affect land produc-tivity. In the delta areasthatareHoodedno more thana meteror so, tworice cropsare typically possible:one ("autumnrice") harvestedearly inthe monsoon,theother ("winter rice") transplantedafter thefloods beginto recede.Jutecompeteswith thefonnercrop. In moreheavily inundatedareas,only a single crop can be grown: either floating rice, able to keepabreastof rising floodwatersto a depthof 5 or 6 meters,or, whereHood-ing is evendeeper,dry-season"summerrice." A singlerice crop is also allthat is possiblein the areasof decayingrivers and poor drainagein thewestern and northwesterndistricts, where long fallow periods havetraditionally beenrequired.Finally, in the hills nearthe easternbordersshifting cultivation is still practiced.1o

Agriculture would be less tied to local conditionsif modernwatercontrolweremore widely available.Masteryof the environmentis still ata relatively primitive level, however.The hazardsof minor year-to-yearvariationsin weatherpatternsand seasonalHooding canbe counteredbyindividual or community effort-building mounds and levees, drillingwells, digging and dredging irrigation channels,elevatingcowshedsonbamboostilts, andso on. But the populationis nearlydefenselessagainstthe recurrentlargervagariesof wind andwater. Realmasteryof the deltawould call for interventionsof a muchlargermagnitude.As yet, however,few large-scaleprojects have been undertaken.I I Ironically, the largesteffort in theregionto modify environmentalconditions,the recentlycom-pletedFarakkaBarrage,damsthe Gangesjust on the Indian side of theborderand hasthe potential to seriouslyinjure the agricultural ecologyof westernBangladesh.12

SettlementandSurvival

We turn next to the humanecologyof Bangladesh.The region'sdemo-graphic history parallels that of the subcontinentas a whole.13 Therecordedpart of this history spansbarelya century,from the first censuscovering Bengal in 1872. Already at that time there were 23 millionpeoplecountedin the areathat is now Bangladesh.Up to aboutthe mid·1920sthe story is one of slow populationgrowth againstenvironmentalandtechnologicalconstraints,interruptedby large-scalesetbacks.14 Afterthat, mortality recededmoreor lesssteadily,a declinehaltedonly by theBengal Famine of 1943 and the disturbancesof Paltition. As a resultpopulationgrowth acceleratedsharply: In the 1930sfor the first time it


exceeded1 percentannuallyover a decade,and recently it has reached2.6 percentper year.

The mortality declineover the last 50 yearshas reducedthe deathratefrom above40 per thousandto a level of about20.15 Various factorshave contributedto this drop, their relative weights not fully agreedupon: fewer famine-relateddeathsas transportand communicationsys-temsdevelopedandgovernmentrelief measuresbecamemoreeffective,16someimprovementsin public health, the introduction of specific treat-ments for epidemic diseases,and lessenedvirulence of some of thediseasesthemselves(notably plague). In causeof death, the most im-portant declines have probably been in the incidence of smallpox,cholera, and malaria-thefirst now eliminated, the others substantiallyundercontrol (althoughmalariashowssomeindicationof resurgence).1TBut in spiteof theseimprovements,the mortality level is still high by con-temporarystandards,evenamongdevelopingcountries.

There has been no comparablesecular trend in fertility. Recentestimatesmostly put the birth rate in the rangeof 45 to 50 per thousand,a rateprobablynot much lower thana centuryago.1S Migration hasbeensporadic,with a large net exodusat the time of Partition and again in1971, but on averagehas little impact on populationgrowth.1!) An esti-matedpopulationtrajectoryand a possiblesequenceof crudebirth ratesfor Bangladeshoverthelastcentury,basedonofficial (Bureauof Statistics)data,are sketchedin Figure 2. The detailedpatternmay be debatable,but thereis little doubtthat the broadfeaturesshownare correct.

Whenwe look morecloselyat thedemographicpicture,we seepopu-lation growth linked tightly to local environmentalconditions.Two pro-cessesare at work. First, wherevernew technologyor resourcesappear,settlementexpandsto take advantageof them. Populationgrowth, as itwere, follows the contoursof local economicopportunity. Second,thepressuresandmomentumof continuedpopulationgrowth sooneror laterexhaustwhatever"slacks" theseopportunitiesprovide; peopleare forcedto seekand exploit whateverother local meansof subsistencethey canfind.

The first processcanbe seenin the expansionof the jute industryinthe late nineteenthcentury.Jutecultivation, concentratedin the central,annually flushed part of the delta, createdan incomethat could financerice purchasesfrom the westerndistricts and later from abroad.Censusfigures since1872 recordslow but steadypopulationgrowth in this corearea, whereas the rest of the country showed comparatively littlechange.2o Migrants flowed in, attracted by the small but sustainedregional surplus; and the marginal improvementin the local economymayalsohaveloweredmortality.

Similarly, the beginningsof significant technologicalchangein agri-culture in the last 10 to 15 years presentanotherkind of oppoltunity.

32 POPULAnON AND DEVELOPMENT IN BANGLADESH

Figure2Outlineof PopulationandVital Rates,Bangladesh,1872-1974

ｾｾ＿ＭＺＰｾ［ＬｾｪｊＺｾｾｾＡ［ｩｾｾｾｭｾｾｭｊＬｾＩｾＧＬｩＧｴｾＺ ;;1 \..'if'

With the increaseduseof irrigation and chemicalfertilizers in the pastdecade,high-yielding, fertilizer-responsiverice varieties have been re-placingtraditionalvarieties,particularlyin the districtsalongthe westernborder (Dinajpur, Rajshahi,Kushtia, andJessore). Theseareasrecordedrelativelyhigh growth ratesin the1961-74intercensalperiod.21

Even small-scaleopportunitiesattract population.Peopleare quickto settleon the char lands in the delta mouth-landthat appearsand asquickly candisappearwith the shifting patternsof siltation.

OppOltunitiesdo not comealong in a steadystream,however,andaspopulationgrowth continuesin any locality, the slacksthey createare

W. Brian Arthur / GeaffreyMcNicoll 33

progressivelyexhausted.But the demographicprocessesat work do not.reverse at that point. Fertility does not adjust downward, nor doesmortality easilyrise again.Migrants oncesettledtend to stay.The resultis a downward pressureon standardsof subsistenceas demographicinertia pressesgrowing numbersagainstthe limits of productivepossibili-ties. Thus we seecultivatedland expandedto its fullest extent,with sig-nificant areasthat are in productionlying in extremelyhazardouszones.Fishing, too, is heavily exploited in inland waters,though less so in theBay of Bengal.The influx of destitutevillagers into the cities-somejustfor ShOltperiodsuntil rural conditionsimproveagain,otherswith nothingto gobackto- is anothermanifestationof this Ricardianprocessof increas-ing marginality.

Thus, not only doespopulationsettlementfollow environmentalop-portunities, but the inbuilt inertia of demographicprocessesexhauststheseover the yearsand pushesenvironmentallimitations as far as theywill stretch.But limitations reachedin the short telm may not hold upover a longer time. Overexploitationcan set into motion long-run forcesof ecologicaldegradationthat are difficult to reverse.An important pastinstance was the building of embankments,where local safety fromflooding was often bought at the cost of acceleratingriver decay overthe long tenn.More recently,populationgrowth in the largestremainingforestareaof Bangladesh,theChittagongHill Tracts,hasbeenshorteningthe fallow periodsin shifting cultivation, leading in tum to serioussoilerosion.And the current technologicalimprovementsin agriculture,en-tailing large" applicationsof fertilizer and pesticidesand greater cropspecialization,may very well havetheir own adverseeffectson the deltaecology, as yet barely visible.2 :!

IndependenceandAfter

With political independencebarelysevenyearsold. wheredoesthe coun-try now standand what are its economicprospects?

The first yearsof independenceweredifficult. Bad harvestsin 1972-73 werefollowed by damagingfloods in thesummerof 1974. Internationaltenns of trade worsenedconsiderably:Prices of foodgrains,petroleum,fertilizer. andcementrosesharply,while jute stagnated.Inflation. fallingreal wages,labor unrest,andpolitical instability addedto the troubles.23

Two yearsof goodharvestsandstablegovernmentput the initial setbacksinto a betterperspective.But evenallowing that it takestime for a newcountry to get propercontrol of its economy,it is clear that Bangladeshfaces difficult problems in the period ahead.

What is the outlook for the industrialsector?The countryhassizablereservesof naturalgas,with explorationstill at an early stage(andbarely


begun offshore). The largest presently known field (Bakhrabad, dis-coveredin 1968) may alone justify constructionof a liquification plant.If so, liquified natural gas could rapidly becomeBangladesh'ssecondlargest source of export earnings, with the possibility before long ofovertakingjute. Naturalgascanalsosupporta fertilizer industry,produc-ing for both domesticuse and export. A growing domesticmarket maystimulateproductionof a wide rangeof consumergoodsand someitemsof capital equipment(agricultural implements, pumps,and so on), andimproving relationswith Pakistanmayreopenthat marketfor "soft" man-ufactures. Perhapseven assembly plants for reexports on the SouthKorea-Taiwanmodel might be developed,although this seemsunlikelyto occur on a scale large enoughto have more than a minor impact.24

In the agriculturalsector,hopesrest on the new high-yielding cropvarietiesand their accompanyingtechnology-inparticular,betterwatercontrol. As is often remarked,rice yields in Bangladeshare only a thirdof Japan's(although they are in line with those of India and most ofSoutheastAsia). A World Bank analysisundertakenin 1911 concludedthat with presentlyknown techniques,togetherwith adequatesuppliesof fertilizer and other inputs and proper irrigation and drainage,riceproductioncould be quadlUpledby the end of the century. Even with-out costly, large-scalewater control projects (i.e., by relying chiefly ontubewells and low-lift pumps), a trebling of output was consideredfeasible.25 But between these possibilities and today's realities standmajor obstacles.First, it is doubtful whetherthe new technologywill beadoptedundercircumstancesat all closeto optimal. Not only mustinputsbe correctlycombinedwith fairly precisetiming over the crop cycle, butthe technologyrequiresa degreeof local infrastructure-credit,extensionservices, distribution, cooperativewater arrangements-noteasily puttogether under presentsocial arrangements.Second,many traditionalcrops,suchas jute andfloating rice, haveasyet no high-yieldingcounter-part. Yield increaseshave occurredmostly in areasthat grow the rela-tively small "summerrice" crop. Third, diffusion of the new technologycannotgo faster than the speedwith which pumps,fertilizer, and otherinputs becomeavailable. Not surprisingly, then, although high-yieldingvarietieshavebeenadoptedrapidly in somedistrictsoverthepastdecade,resultson a national level have so far beendisappointing.26 Writing in1975, NoazeshAhmed (1976, p. 19) concludedthat "irrigation facilitiesandthe input packageprogrammesso far extendedhavehad little effecton the national productionof rice. The higher rice production in thecountrystill dependsmostly on good weatherconditions."

In this brief accountwe haveseenthat Bangladeshat independenceinheriteda largely static and traditional rural economywith a small in-dustrialbase.A denselysettledpopulation,growing rapidly over the lastfew decades,meshestightly with a hazardousnaturalenvironment.The

W. Brian Arthur / GeoffreyMcNicoli 35

local administrativesystem is weak, limiting the national govemment'sability to influence eventsat the local level. Importantopportunitiesforindustrial growth exist, but becauseof the economicdominanceof agri-culture, any hope of rapid changenecessarilyrests on progressin therural sector. Here we have a recognizedtechnologicalpath to follow,althoughone that has severeobstaclesalong the way.

The Local Contextof Development

Underlying the rural economy and governingmuch of its performanceis a complex local social system.In this section,we ask how this systemworks and how it is supported,and examinethe ways in which it pro-motes or impedesvarious kinds of social change.In shifting our focusfrom the national to the local level, we are led into territory equally un-familiar to the demographerandthe developmentplanner,but an areainwhich both the economicanddemographicfuture of Bangladeshis likelyto be decided.

Dynamicsof Landholding

Local organizationin rural Bangladeshis closely tied to the ownershipofland. Land, in a countrywith few otherproductiveassets,bestowsstatus,power,and aboveall security.Although in good yearsincomefrom wagelabor can supporta family, over longer periods,with a saturatedlabormarket,it is a poor guarantee.Only ownershipof land, or usagerights toit, prOvide a long-run surenessof accessto the local social product. Notsurprisingly, then, wealth and social standingin lUral Bangladesharemeasuredin termsof landholding.27

With virtually no farms that are large in absoluteterms,small varia-tions in landholdingcan lead to markeddifferencesin social class.Divi-sions are somewhatarbitrary, but we might distinguish four classesofagricultural families.28 Thosewho control about 3 hectaresor more con-stitute the rural elite-5 percentof families with roughly 20 percentofthe agricultural land. Theseare the landlordsand the surplus peasants.Beneaththem arethe middle peasants,thosewith about1 to 3 hectares-roughly 25 percentof families in agriculture. Many of these are bothlandownersandtenants.Next in standingare marginalpeasantswith lessthan 1 hectare,including the pooresttenant farmers and sharecroppers.Finally, at the bottom of the scale are landless agricultural workers,forced to rely on agriculturalwage labor and whateverother sourcesofincome they can find. These last two categories covermore than two-thirds of the agriculturalpopulation.Subsistencefor them in all but the


best of times dependson relations with the larger landownersto helpthem securewage employment.

Declining landholding lies behind much of the present rural im-poverishment.The actual dynamicsof landownershipare complicated,but the main processesare closely tied to population growth and areworth looking at in somedetail.

If populationwerestationary,landcould be turnedover from fatherto son in a fairly constantpatternover the generations.Where popula-tion is growing rapidly, as in Bangladesh,patternsmust change,andthere are a limited number of possibilities for accommodation:moreland broughtundercultivation;more falms on land alreadyin use; morepeopleon existingfarms; or more peoplelandless.Eachof theseoptionsappearsto havebeentakenin Bangladesh.

Historical dataon landholdingare hard to come by, but it appearsthat in the early part of the centurythe growing populationwas accom-modatedby expansionof cultivated land and by an increasein thenumberof fanns (as land was increasinglysubdividedby inheritance).In the last 20 to 30 yearsthesetwo possibilities have largely been ex-hausted.By the early 1950sexpansionof cultivated land had virtuallyceased;and as fann sizesgrew smaller through constantdivision, hold-ings were increasinglypushedbelow subsistencelevel-the size neededto support a family's consumption.29 A limit to the multiplication ofsmallerandsmallerfarms was in effect reached.ｂ･ｬｾｷ a certainsubsist-encesize,families werelikely to fall into debtin a poorseason,enteringadownwardspiralthatfrequentlyendedwith the lossof theirholdingsto themoreaffluent. Populationgrowth, no longerable to be accommodatedonincreasingnumbersof smaller fanns, addedinsteadto the landlessandto the numbersper fann.

The figu!es bearout this sequence.As Table 2 indicates,numbersofsmallerfannshavenot increasedin the last20 years;in fact, the sizedis-tribution has been ahnostconstant.30 Landlessness,on the other hand,has greatly increased.The trend is strikingly evident in recentsurveys,which show the landlessagricultural populationgrowing from about 15to 20 percentin the early 1960sto between30 and 35 percenta decadelater.31 Simultaneously,the averagenumbersper fann haverisen aswell.During the 1961-74 intercensalperiod, the increasein the agriculturallaborforce wasroughly5 million andthat of landlessagriculturalworkersperhaps3.5to 4 million, leavingmorethanamillion workers,togetherwiththeir dependents,addedto the populationon existingfanns.32

The implicationsof theseprocessesare several.Obviously, facedbythe prospectof losing land, individual families have less security. Thethreatof landlessnesshangsover the greatmajority of landholdingfami-lies. Purchasingland or holding on to what onehasbecomesan overrid-ing concern.


Also, the increasingnumbersof peopleper fann createpressuretoseekothersourcesof income.Detailedstudiesof individual fann familiesrevealan astonishingrangeof income-generatingactivities in addition tofanning. One analysisof a village economyin a national-incomeformatfound less than half of village income attributable to cultivation, andonly two-thirds to all of agriculture.a3 For the poorerpeasantspetty trad-ing, fishing, and handicraftscan supplementwage labor. The better-offare often active in such small businessesas marketingagricultural prod-ucts, distributing inputs, and moneylending.34

Table2Distributionof FarmAreaby NumberandSizeof Farms(OperationalHoldings),Bangladesh,1960,1968,and1974

Percent of Farms Percent of Farm Area OperatedRanked by Size(Quintiles) 1960 1968 1974

ｆｩｾｴ (bottom 20% of farms) 2.3 3.0 1.6Second 8.4 8.3 3.4Third 14.3 13.3 14.4Fourth 21.8 22.4 25.6Fifth (top 20% of farms) 53.2 53.0 55.0

SOURCE: Alamgir (1975), p. 268. Data compiled from 1960 Agricultural Census, 1968Master Survey of Agriculture, and village surveys taken in 1974 by Bangladesh Instituteof Development Studies.

As a further consequence,growing numbers of landless put anincreasing pressureon the already strained labor market. We notedearlier that real wages have fallen substantiallyin recent years. Thistrendcouldbeamelioratedsomewhatby the new agriculturaltechnology,which is modestlylabor absorbing,but offsetting this hope is the growthof the farm population itself. Farmersare more likely to employ theirown family membersbefore they hire others.ali Finally, as people arepushedoff farms, out of ownershipinto wage labor, agriculture is slowlybecomingcommercialized.Small-scalepeasantownershipand localizedlabor marketsare giving way to an economybasedon impersonalem-ployer-employeerelationships,a primary sourceof changewe will lookat shortly.a8

Social Organizationin a Changing Society

The precedingpicture of a skeweddistribution of accessto the ruraleconomyis a major part of the Bangladeshreality. But accessis not so


simply determined.As in any traditional society,ties of kinship, patron-age,andneighborhoodmaycut acrosslandholding,servingto strengthenor weakenan individual's claim on the social productand to modify hissenseof risk or security.

The patternof rural socialorganizationat the local level is intricate.Peoplescatteredthrough varioushorizontallayersof landholdingclassesare boundtogether,sometimesstrongly, sometimescasually,into group-ings of severaldifferent kinds. Table3 gives a schematicsumma.ryof themoreimportantof these,achievingcomparativesimplicity of descriptionat the cost of blurring many fine distinctions and ignoring local varia-tions.37 Besidesthose groupings mentioned, there are divisions alongreligious lines (the Hindu minority is separatein many respects),byoccupation (the residue of caste,with modem status rankings super-imposed),and by length of residencein a particular locality. It is suffi-cient here,however,to pick out the more poweIful organizingforces.

Kinship ties at severallevels enmeshindividuals in a. rangeof obli-gationsand in tum help to distributethe burdenof risk. The nuclearorpatrilineally extendedfamily is the basic social and economicunit ofBangladeshsociety.Beyondhis immediatehousehold,a personhas well-defined duties to his bari (homestead)and somewhatweaker ties tolarger, kin-basedgroups (paribar and gusthi ). For example, the barioften operatesas a corporateentity with land held in the name of itshead, who exercisespatriarchal control over members.38 Paribar andgusthitend to dominatein suchmattersas selectionof spouseand nego-tiation of dowry or bride-price,the upbringingof children, and disposalof assets.Theseextendedfamily and lineage groupshave more signifi-canceamongthe bigger landowners.For families with no land or only ahousecompound,thereis little economicrationalefor emphasizingrela-tions with closebut equally impoverishedkin.

For families with a tenuousplace in the economy, alliances withleading surplus fanners in patronagegroupings (shamai, reyai) havemuchto offer. In exchangefor allegiancethe small landholderor landlessworker may receivepreferentialemployment,somesupportin bad years,seasonalcredit, andotherbenefits.The leaderin turn obtainsa followingto supporthim in local electionsand disputesand, obviously, a senseofimportance.The nucleusof a patronagegroup consistsof the leaderandhis close relatives; wider membership typically comes from adjacenthouseholdswithin the village. (A village may contain several suchgroups.) The functions of patronagegroups vary somewhatdistrict bydistrict. In somethey are purely political; in others quasi-official func-tions-partceremonial,part judicial-arealso present.Membershipmaychangeover time, and not all families necessarilybelongto any group ofthis sort. In at least somevillages there are also informal factional divi-sions separatefrom shamaj groupings (or occasionallywithin them).

w. Brian Arthur / GeoDrevMcNicoll 39

Thesefactions (dal), often at odds with eachother, have a more fluid,even shadowypresencethat is not openly discussedwith outsiders,butmay form an importantpart of local political reality.39

Natural villages or hamlets,the basisof rural settlementand localallegiancein much of Asia, are weak in Bangladesh-afact that has im-portant consequencesfor development.The natural village (gram )issocially defined, and residentshave a clear perceptionof its territorialboundaries;but these units tend to have no corporatefeatures, littlecohesiveidentity, and only a residualdegreeof solidarity.40 One writerhas observed:"A man'sduties are, in order, to his own family (bari),then toward his paribar, then to his gusthi and then to his village."41Settlementpatternsare partly responsible:In land subject to extensiveseasonalflooding, small groups of homesteadsare clusteredon raisedground (often moundsor levees)built up from surroundingflood plains;hence,dispersedor linear settlementsrather than nucleatedvillages arethe rule.42 But the relative lack of function of villages in the societycanalso be tracedto the colonial failure to provide effective local adminis-tration. As a consequence,village life is segmented.A man may residein one village, attenda mosquein another,patronizea marketin a third,and cultivate plots of land in any or all of them. For adjudication ofminor disputeshe may calIon the headof his gusthi or on the leaderofthe shamajto which he belongs;for assistancein ploughingor harvestinghe may turn to othermembersof his paribaror to wage labor from dis-tant villages.

Territoriality is not completelymissingfrom Bangladeshsocial orga-nization, however. The subvillage neighborhoodcluster, the para, issometimesa cohesivesocial unit, although it has little explicit role atpresent.And as we have seen,patronagegroupingsalso have a kind ofterritorial basis,albeit a fluid one.But for the most part, functionsusuallyascribedto the village community in peasantsociety are filled in ruralBangladeshby a variety of nonresidentialand overlappinggroupswithmoreor lessspecializedconcerns-religious,political, economic,andsoon.

From this brief description,kinship and patronageties standout asthemostpowerful organizingforcesin rural society.Thesetwo traditionalforces togethercharacterizethe local system.They act as a simple, non-market distributive mechanism,channelingaccessand security down-ward; and,by binding togetherpeopleof varioussocial levels, they havetendedto diffuse any strongmanifestationof class.They havegiven thesystema high measureof stability.

Under the twin pressuresof populationgrowth and increasingcom-mercialization, there are signs that both kinship and patronagebondsare weakening.43 As population grows, the kinship-patronagedistribu-tional system may becomeoverloaded:Assets are diluted through thesubdivisionof land,supporters'needsgrow, andnumbersof peopleon the

Table"Elementsof RuralSocialOrganizationin Bangladesh

Grouping

chula,khana,ghar(household)

bari (homestead)

paribar (lineage)

gusthi (clan,kin group)

OescriptionandBasisof Affiliation

Nuclearor jointfamily

Extendedfamily com-prisingseveralnuclearor joint householdssharinga homecompound

Groupof bariswhosemembersareagnaticallyrelated;would rarelyoverlapa village

Kin based,but moreinclusivethan paribar;sometimesextendsbeyondVillage boundary

ApproximateAverageSizeor SizeRange

5-6 persons

4-6 households(20-30persons)

50-150persons

Wide rangeofsizes

MajorFunctions

Main productiveand-reproductiveunit,but often operatesasclementof bari orparibar

May operateaseconomicunit

Haswide-rangingcontrolovermembers'behavior(e.g.,mar-riage,dispositionofproperty)

Control functionsaresimilar to paribar;headsof constituentparibarsmayactasinformal localleadership

AssessmentofStrengthor Cohesion

Strong

Typically strongandcoheSive

Strong

Often strong

Table 3 (continued)

GroupingDescriptionand Basis

of AffiliationApproximateAverage

Sizeor SizeRangeMajor

FunctionsAssessmentof

Strengthor Cohesion---------------------------------para,kandi(neighborhoodgroup)

gram (local village,hamlet)

milat (mosquecongregation)

shamai,reyai(patronagegroup)

Clustersof adjacenthomesteadsor householdswithin a village; memberslikely to be all of onereligion

Sociallydefinedvillage;often not coincidingwiththeadministrativeiyrecognizedunit

Groupdefinedby attend-anceat samemosque;often drawn from morethanonevillage

Definedchiefly by patron-client bonds,usuallycen-teringarounda dominantleaderor line.lgegroup;typically membersliveadjacently

100-200persons

c. 100households(400-800persons)

Wide rang,

Wide rangeof sizes,from a few householdsto an entirevillage;membershipwelldefinedbut fairlyfluid

NU/Il:

few and residual;rarely hasanycorporatefeatures

Chiefly ritual andceremonial

Mobilization ofpolitical resources;interfacingwithgovernmentadminis-trativeanddistribu-tive structures

Moderatelyweak,unlesscoincidingwith a shamajor lineagegroup

Weak

Weak

Often strong

---_._----------------------

SOURCES:Seenote37.


surplus fanns rise. Fonner clients must be abandoned,left to attachthemselvesto the fonnal wage system.And growing commercialization,reflecting the new valuesand opportunitiesassociatedwith technologicalchangeand increasedurban contact,is beginningto transfonnan econ-omy basedon personalrelations into one that is fonnalized and mone-tized. This process,which might be seenas a healthysign of moderniza-tion in many countries,is a mixed blessingin Bangladesh.Fannersfindoutletsotherthantraditionalpowermaintenancefor their surplus: invest-ment in nonagriculturalenterpriseor in education,new consumergoods,and so on. Increasingly,there are signs that large landownersare reluc-tant to let to their kin; sharecroppingis giving way to pure tenancy;and,as noted earlier, patron-clientbonds are being supplantedby an em-ployer--employeerelationship. Market forces that pennit the wealthierpeasantsto shakefree from their fonner obligationstend to erodewhatlittle security the landlessand near landlessnow possess.More peoplein theselatter classesarecompelledto be geographicallymobile, respond-ing to seasonallabor demandsand in turn depressinglocal wagerates.H

Although traditional patronageandkin ties probablystill dominateruralrelationsin mostof the country (systematicinfonnatlonon the paceandextent of changeis lacking), the directionsand impetus for movementtoward a more fonnalized, impersonaleconomy are clearly evident.45

Our overall impressionof rural social organizationin Bangladesh,then, is one of diffuseness.Duties and obligationsrun in various direc-tions, and functions are split amongdifferent kinds of social groupings.Local society is fragmentedinto groupsorganizedaround leading fami-lies, which are often at odds' with one another. No strong territorialgroupings exist to pull community interestsinto line. This traditionalstructureis changingslowly, increasinglycoming to be basedon fonnalcommercialrelations,but this processis still in its early stages.We turnnow to the implicationsof this local systemfor rural development.

The Politics of Local Change

We have lookedbriefly at the patternsof economicand social organiza-tion in rural Bangladesh.How do theserelationsset the contextof ruraldevelopment-definingcommand over resources,limiting the scopeofindividual action, andpromotingor hinderingeconomicprogress?

Agriculture in Bangladesh,as we have seen,has yet to show thedramatic improvementsthat the new technology can potentially offer.Yields are still low, and governmentattemptsto foster changehave metwith mixed success.Certainly such problemsare by no meansunique,but in Bangladeshtwo specialcircumstancesworsenmatters: the generalsocial diffusenessjust describedand the absenceof the kind of local ad-ministrative system that could prOvide a finn institutional setting fordevelopment.

W. Brian Arthur / Geaf]reyMcNicoll 43

Socialdiffuseness-thelack of communityintegrationandpurposewenotedearlier-wouldbe lessof a problem if peasantfanning were carriedout on an individual basis. But the new agricultural technologymakesgreaterdemandsthan before on organizationand cooperation,both ofwhich are weak in rural Bangladesh.Village society was seen as cutacross by various social groupings, often in conflict. In the case ofrival factions,competitioncan be outrightly violent. Bertocci (1970) hasdrawn a striking and disturbing parallel between the emergenceandpotencyof factional groupingsin Bangladeshsocietyand the situation inSicily that led to the growth of the Mafia. Vertical relations can alsoinvolve violence and intimidation: A recent seriesof village studiesbyShapanAdnanpresentsa starkpictureof this aspectof village life. 46 Thisatmosphereof competition and conflict not only provides an insecuresettingfor individual initiative, but also ensuresthat the benefitsof devel-opmentgo mostly to the powerful. Distribution of fertilizer and controlof irrigation areoftencorneredby a dominantfamily or factional group,to the detrimentof the restof the community.

The problem here, of course, is not that Bangladeshvillagers areinherently unableto cooperate-manyefforts to fonn rural cooperativeshave beenquite successful-butrather that, without strong outsidesup-port, traditional social and economic segmentationeventually blockscollaborativeeffort. In somesocieties,territorial bondssupplya sufficientmeasureof community responsibility to avoid such an outcome,but inBangladesh,as we haveseen,territorial affiliation is weak.

H traditional social arrangementsimpede agricultural development,the local administrativesystem imposed on them does little to helpmatters.In Bangladesh,as we notedearlier, the various colonial regimesdid not leave behind a strong system of local governmentcapableofrespondingto nationalgoalsand providing a finn institutional settingforrural change.And more recent attempts to fill this gap have merelystrengthenedthe existingobstacles.The basicunit in theregionaladminis-trativehierarchyis the union,correspondingto about15 censusvillagesonaverage,with a populationof 15,000to 20,OOO.4i Unions are governedbyelectedcouncils,which havebroadauthority for local welfare and powerto levy land taxes. When thesecouncils were instituted under universaladult franchisein 1959, most members,not surprisingly, camefrom therural elite-the larger surplusfannersand the leadersof dominant line-agesor patronagegroups.48Council membershipand, in particular, thechainnanshipwere major new sourcesof local power and patronage.Tothe extent it had wished to do so, the national governmentin reachingdownwardthroughthe administrativehierarchyfailed to gain a purchaseat the union level. Far from introdUcing a territorially basedstructureatodds with traditional village power relations,the administrativesystemat its lowest level becameeffectively an instrument of those relations.

As a resultof this captureof the local systemby the rural elite, their


interests,not those of the averagefanner or those proclaimedby thenationalgovernment,continuedto dominatethe rural scene.Under thetraditional social systemtheseinterestshad little to do with agriculturalinnovation. For the rural well-off, lending money at high interestrates(oftenmore than100percent),buying up mortgagedland, and buildingthestrengthof followers providedfasterroutesto positionandstatusthanpainstakinginvestmentin agriculture.40

H attemptsto createan effective administrativestructureindepen-dent of the local elite thus foundered,most governmentprogramstodirectly promote rural developmenthave met a similar fate. The bestknowncaseis the history of the cooperativeschemessetup by the Bang-ladeshAcademy for Rural Development(the Comilla cooperatives).liOIn their own tenns, thesecooperativeshad substantialearly successinorganizingsmall and middle peasantsand in providing memberswithseed,fertilizer, credit, and so on. But Over time, and especiallywith theattemptedcountrywidereplicationas the IntegratedRural DevelopmentProgramme,mostcooperativescameto be dominatedby the rich fannerswho could get themselveselected to the managingcommittees. (Theprocessis of coursefamiliar in cooperativemovementsin many coun-tries.) Theserich farmerswere able to monopolizeinputs, re-Iendcheapgovernmentloansto othersat higher rates,andcontrol irrigation groupsto the benefit of themselvesand their supporters.The original intent ofthe scheme,to mobilize and assistthe small fanner, was achievedonlybriefly. Programsthat have sought to benefit the poorestrural classes-the Rural Public Works Programmeand various other food-for-workschemes-seemto have been less open to direct capture. But if theseefforts had beenon a scaleto makea significantimpacton rural employ-ment or distribution, we would expectthat before long they also wouldhave confonnedto local political and administrativerealities.51

In short, the segmentedsocialorder that characterizesrural Bangla-desh,while it perfonnssomeusefulfunctions,works to obstructcollectiveagriculturalinterestsat the village level. Governmentattemptsto imposeoutsideorganizationanddirectionhavefor the mostpartbeentakenoverby traditional forces, which havedivertedthem to their own ends.

As local societybecomesmorecommercialized,thesepolitics of ruraldevelopmentarelikely to change.Rural power,traditionallyderivedfromlineagerights andmorerecentlyfrom local governmentauthority, is nowshifting to an economicbase.Increasingly,thosewho cancapturea dis-proportionateshareof the new technologyand its supportinginfrastruc-turebecomethenewrural elite. Undercapitalistagriculturethe landlords,as Wood (1976, p. 149) puts it, "exchangethe benefits of a respectedleadershippositionfor thosebasedon themorenakedreality of economicpower." It is possible,of course,that this new economicbasisof societywill transfonntraditional patronageleadersinto a classof well-behavedpetty capitalistsanxious to invest in rural growth. But if the new elite

W. Brian A,..:!lw ! ［ｾ ..:::.:;Jrc'l :VicSicoil 45

doesbeginto investheavily in high-yieldingtechnology,thereis a dangerthat this could result in a speedingup of land transferand a further con-centration of resources,with negative consequencesfor the spreadofbenefitsto the lower levels of society. Insteadof being a powedulforcein the modernizingof society, high-yielding agriculture could well be-come little more than the new instrumentby which a rural elite main-tains its traditional position.ｾＲ

This picture of local societywould not be completewithout mentionof the urban scenein Bangladesh.Cities presentlyhave little weight inthe overall economicandsocialstructure,containinglessthan 10 percentof the population. Dacca (1974 population 1.7 million), Chittagong(900,000),Khulna (400,000),and Narayanganj(270,000,but essentiallypart of the Daccametropolitanarea) are the only major centers.ｾＳ Butwith this small baseit takesonly a slight rural-urbandrift to inducerapidurban growth. Thesefour cities nearly trebled in size between1961 and1974-anannualrate of increaseof 8 percent,or more than three timesthe growth rate of the rural population.Yet even at such a growth rateurban areasare absorbinglessthan a fifth of the rural natural increaseeachyear.

New arrivals in the cities come from the full rangeof rural society.For somemigrants,mostly from the rural elite, the cities are openingstoan urban middle class, offering educationand the prospectof jobs ingovernmentor theprivatesector.For the lesswell-off, this small modemsectorsupportsconsiderablelower-statusemployment,rangingfrom wagelabor in manufacturingand constructionto fringe activity in small-scalecommerceandservices.And in badseasons,suchasthatof 1974,the citiesalso becomethe last refuge of the rural destitute.

The recencyof significant urbanizationin Bangladeshmeansthat itis too early to assessthe natureof the societythat is evolving. Almost bydefinition urban behavior is less inlluenced by family and kinship ties,whether becauseof emergingmiddle-classattitudes or becauseof theimperativesof day-to-daysubsistence.But new social groupingsappear.The poor now tend to be residentiallyseparated,often in squattersettle-ments (Daccawas estimatedto have 400,000squattersin ＱＹＷＲｾＴＩＬ andawarenessof their commoninterestsleadsto somecommunitysolidarity.Young people,particularlystudents,becomea distinctivesocialgroupandsometimes,as in the eventsleadingup to the 1971 war, a political force.And factory workers and even governmentemployeesare mobilized inlabor unions. Administratively, a sometimestenuous stability is main-tainedwith help from a visible military and police presence.

To summarizethis accountof the local contextof development,weseeaccessto land anddistribution of securityas the important traditionalorganizing forces in the rural economy and society. But both arethreatenedas landlessnessincreasesthrough rapid population growth.


Agricultural changeis impededby a social systemthat allows develop-ment initiatives to be capturedby the better-off, and social cohesionisfragile and exclusive,making it difficult to organizechangethat wouldbenefitsocietyas a whole. A growingcommercialismmay help "modern-ize" this traditionalsociety,but at the sametime it is weakeninginformalmechanismsof distribution through patronage,and may well lead to afurther concentrationof assetsso that the larger massesare not reachedby development.Finally, cities are growing rapidly, but are still small,providing at besta modestoutlet for the rural poor.

Demographic Processesin Rural Bangladesh

We havedescribedthe local settingin which individual and family lifein Bangladeshis played out and have seensome of the changesnowtaking place.A major factor in thesechanges,aswe saw also at the ag-gregatelevel, is the rapid paceof population growth. This puts ever-increasingpressureon a restrictedanduncertainenvironment,makestheswift adoptionof newproductivetechnologiescritical, and,with a limitedland area, causesa downward spiral of-land impoverishmentwherebythe security of one'splace in the distributional systembecomesall im-portant.In this sectionwe examinethe demographicprocessesthat driveBangladesh'spopulationgrowth, settingour analysiswithin the contextof the family. We begin by looking at mortality: how the falling deathrate has altered the family life cycle, how mortality processeswork atdifferent levels of society, and the prospectsfor continueddeclines. Indiscussingfertility the relevantcontext is not only the family but alsothe distinctive place that women occupy within it in Bangladesh.Aftersettingthis contextwe then go on to examinethe actualmechanismsoffertility in somedetail: the biological constraintson numbersand timingof births and the particularsocial andeconomicincentivesthat maintainfertility at high levels.

TheFamilyunderChangingMortality

In the pastthreeor four decadesdeathrateshavehalved in Bangladesh.For thefamily this meansseveralchanges:More childrensurvive,parentslive longer, child dependencymay haveincreased,more sonssurvive toinherit land, andso on. We look at someof thesechangesnow.

Most peoplebelongto eithernuclearor pahilinealextendedfamilies,the latterusuallyorganizedasseparatenuclearfamilies groupedtogetherresidentiallyasa homestead.Families,of course,havea life cycle, so that

W. Brian Arthur I GeoffreyMcNicoli 47

at any time a substantialproportionof the populationwill be found livingsingly or in othertypesof householdsthan the nuclearor patrilineal type.Table 4 indicates that in 1960 one-third of all householdswere "non-standard"in this sense.Given Bangladesh'salmost universal marriage,the situationlargely reflectsdemiseof families throughdeathof a spouse,usually the husband.

Table4Distribution of Bangladesh Householdsby Family Composition, 1960

Compositionof Household

One person onlyHusband and wife onlyHusband and/or wife with own childrenHusband and/or wife with or without

own children but with parents and/ordaughter-in-law

Households comprising other relativesor nonrelatives

All households

Distribution of Households (Percent)

Urban Areas Rural Areas Total

11 4 55 5 5

29 33 33

22 31 30

33 27 27

100 100 100

SOURCE: 1960 Housing Census (Census of Pakistan 1961, vol. 9, Table B).

The family life cycle startswith marriage.For the averagewomanin Bangladeshthis takesplacebeforeage15. By age20 only 5 percentofwomen are still single; less than 0.1 percentnever marry. Men tend tomarry in their 20s; the averageageis 23, but with a bigger spreadin agethan for women.Theremay havebeena slight increasein marriageagesin recentdecades,particularly in the last 10 to 15 years; later we con-sider the prospectfor more substantialchangesin the future.r.5 As wenotedearlier there is somedoubt about the exact level of fertility, but atotal fertility rateof aboutsevenchildren per womanmay well prevail.56

What has been the impact of mortality decline on family size? Arough answercanbe obtainedby contrastingthe processof family build-ing under the presentmortality regime with that of the 1930s,assuming(realistically) that fertility patternshaveremainedmuch the same.Sup-posethat in both casesa womanis marriedat age 15 to a man aged23andsurvivesthrough her reproductiveyears. Figure 3 shows the result-ing size and durationof the nuclearfamily in the two mortality regimes.The family is assumedto breakup with the deathof the husband,whoselife expectancyat marriagewould havebeenabout30 yearsin the 19305,40 yearsin the 19705.(Figure3 doesnot give any indicationof variability


Figure 3AverageSizeandDuration(to Deathof Husband>of NuclearFamily in Bangladesh,underMorality Regimesof 19305and19705

in mortality andfertility amongfamilies. For example,the poorestfami-lies maystill experiencetheolderpatternof mortality. Someof this varia-tion we discussbelow; herewe dealwith the average.)

The main aspectsof the new demographicsituationnow begin toemerge.Morechildrensurviveatall marriagedurations:Thenewaveragefamily is bigger.In the 1930sonly abouthalf of all childrenborn reachedadulthood; in the presentdecadethree-quartersdo. (See Table 5 for


approximateproportionsof childrensurviving throughchildhoodyears.)Themalepatriarch,living ten yearslongerthan previously,lengthenshisperiodof dominationoverhis family andcausesa correspondinglylongerwait for his children to inherit. Land must now be <;livided among anaverageof 2.4 surviving sonsinsteadof 2.0, assumingthat daughtersarebypassed.And with improving longevity for womenas well as men, theexpectedperiod of widowhood remains lengthy (around ten years),a fact thatbecomessignificantwhenwe considerthe dependentpositionof women.

TableSApproximateAge Patternof Child Mortality,Bangladesh,19305, 19505,and19705(ExcludingEffectsof Major Disasters)

Age in Years

Probability of Surviving to Each Age (Percent)

1930s 1950s 1970s

125

1015

7265585552

8075706765

8581787675

SOURCES: Estimates for the 1930s and 1950s assume life expectancies of 28 and 38 yearsrespectively, distributed according to the Coale-Demeny West model pattern. Estimatesfor the 1970s are taken from the life table derived from the 1974 Retrospective Survey.

The fact that more children survive than previously might seemtoimply that the burdenof child dependencywithin the family has risensubstantiallyin recentyears.Yet this is not necessarilyso. Child depend-ency in Bangladeshis difficult to define. Childrenare usuallyput to use-ful work betweenthe agesof six and ten, and by their early teenstheyhaveceasedto be economicallydependent.57 Moreover,in the early yearsof marriage,when the dependencyburdenis greatest,the nuclearfamilyis quite likely to be part of a larger patrilineal family group. Most ofall, though, marriages(survival of both partners) in this new situationlast about ten years longer. Thus, although more children survive toburdeneachfamily, the family has a longer productivelife in which toreap the benefitsof thesechildren. The higher dependencywithin thepopulationis scarcelynoticeablewithin eachfamily.

We canconcludethatoncethe family is setup it doesnot experiencethe resultsof the mortality decline very directly in its life cycle. Where,then, are thesedemographicpressuresfelt? At one point in the life cyclethe economicsof continuedgrowthareexperienceddirectly-in settingupa household,when landmustbe subdividedor new meansof st:stenance


found. Even here,though,the way is likely to be smoothedby the patri-lineal family. Many of thesefamilies, aswe haveseen,act to somedegreeascorporateunits-sonscanfarm patrilineal land, with the fatherholdingtitle until his death.To a lesserextentdemographicpressuresarealsofeltwhen the family tries to enter a child into the stretchedlabor market.Oncea shareof economicresourcesis procured,however,the family cansettledownto alife thatwould appearto belesssteadilyerodingthantheaggregatefigureswould show.

Mortality:ProcessesandProspects

While family life haslengthened,reflectingcurrentlower mortality, qual-ity of healthin Bangladeshremainspoor. Per capitaprotein and calorieconsumption,alreadyat very low levels in the early 1960s,are estimatedto havedeclinedover 1960-75.The deteriorationwas probablymore pro-nouncedfor the landlessrural population and the urban poor, both ofwhom experienceda substantialdrop in real wagesin this period.58 ChenandChaudhury(1975) think it likely that averagebody weight declinedin the1960s.Malnutrition impedesboth physicalandmentaldevelopmentof theyoungand,throughits weakeningof the body'sdefensesystems,isa basiccontributorto mOltality andmorbidity in thepopulationasa whole.

In the conventionalcause-of-deathcategories,the largestcontribu-tion to presentoverall mortality is from diarrheal diseases(includingsomecholera),which may now accountfor as much as a third of alldeaths.Other significant contributorsare tuberculosisand other respira-tory diseases,tetanus(an importantcauseof neonatalmortality), andtheusual chronic diseasesof old age, which have becomemore salient asepidemicdiseasesare reduced.59

Thereis little direct evidenceon the prevalenceof illnessin the pop-ulation, but it is undoubtedlyhigh. Public knowledgeof diseasetrans-mission is primitive and healthservicesare sparse.60 For even moderatedeathratesfrom infectiousand parasiticdiseasesto persist,a large pro-portion of the populationmust suffer from thesediseasesandexperiencethe enervationand low moralethat go with them.

Likely future trends in mortality are difficult to assess.Mortalityrisks, far from being equally distributedthrough the population,closelyreflect accessto the local economy.To simplify matterssomewhat,twomortality processescanbe seenat work, onefor thoseabovesubsistence,the otherfor thosethrown uponthe margins.For thoseabovesubsistence,muchwill dependon further progressin public health. Relatively easytreatmentfor diarrheal diseasesnow exists through the prevention ofdehydration (oral rehydration therapy), which has yet to be widelyexploited. Greater availability of antibiotics and advancesin general


healtheducationwould alsohavepotentiallyimportanteffectson mortal-ity. All of these,however,call for substantialimprovementsin the level ofhealthservices.For the above-subsistencedeathrate to continueto fall,developmentof an effective health infrastructureis probably more im-portant than dramaticimprovementsin the economy.

Those compelledto subsistat the margins-whethergeographicoreconomic-facemuchhighermortality risks. On the marginsof settlement-for examplethe char lands in the delta mouth-recurrentflooding andtidal borescausehigh disasterfatalities (seeTable 1). For thoseforcedonto the economicmargins-anumericallymore impOitant group-alessdramaticprocessoccurs.Thesefamilies becomegradually cut off fromthe distributional system,and in hard yearsare the first to die. Figuresarescant.Datafrom an intensivestudyof onesmall rural locality indicatechild mortality ratesfor the poor nearlydoublethoseof the well-off, andan even greater class differential in the frequency of widowhood inmiddle life. 61 Another village study, giving figures from a famine year,showsdeathratesin landlessfamilies threeto four timesthoseof familieswith more than1.2 hectares(seeTable6).

Table6DeathRatesby Family Landholdingin a FamineYear,CompaniganjThana,Noakhali District, 1975

Size of Landholding(Hectares)

oLess than 0.20.2-1.2More than 1.2

SOURCE: McCord (1976).

Crude Death Rate(Per 1,000)

35.828.421.512.2

Death Rate ofChildren Aged

1-4(Per 1,000)

86.548.249.117.5

Overall, therefore,we seea bifurcatedmortality processin Bangla-desh: lower risks for thosewho are well-off, muchhigher risks for thoseon the margins.With growing impoverishment,this secondtype of mor-tality could becomemore dominant. Rough calculationsshow that theprocessof increasingmarginality through landlessnesscould add two tothree points to the averagecrude death rate in the next decade.Butpreoccupationwith averageratesmakeslittle sensewhenmortality variesso widely by class: Improvementsin thenationalaveragewould be smallcomfort to the manywho risk being pusheddownwardin the economicscaleto face a distinctly harsherregime of mortality.


Womenandthe Family

Before discussingfertility it is worth pausing to look at the place ofwomenin Bangladesh.This subjectscarcelyarosein our earliersurveyofeconomicandsocialorganization,for in Bangladeshasin few othercoun-tries women'swork anq social life lie within the householdand theirstatusis conferredby husbandor son.Very few womenwork in the fieldsor in other jobs away from the house.For the most part, activities inrural areasare clearly divided betweenthe sexes.Men undertaketheoutsidestagesof grain production-sowing,crop husbandry,andharvest-ing; womentakeon the stagesthat canconvenientlybe pedormedwithinthe householdcompound-husking,parboiling, andpreparingnext year'scrop. Women also fulfill the usual rural householdduties: care of chil-dren,preparationof meals,andpossiblycareof a few domesticanimals.62

We have alreadyarguedthat the main problem for the individualman in a marginal environmentis to tie himself into the distributionalsystem-topurchasesecurity through any available economic or socialmeans.Women are more constrainedin their searchfor security. Indi-vidual initiative is largely ruled out. Strong traditionsof purdah dictatethat respectablewomendo not engagein tradeor fieldwork or leavethefamily for other than traditionally specifiedvisits to relatives.Separationand divorce are real threats,especiallyif the womanprovesto be child-less.83 And early loss of the husband'sability to provide, through illnessor death,canbecomea disasterfor a family in its middle years.84 Thesethreatsof loss, togetherwith social restrictionson acting for themselves,put womenin a positionof high risk that risesas life progresses.Womenrespondby strengtheningwhat familial bondsof security they can. Thepart inheritanceallowed to daughtersunder Islamic law (half that towhich a sonis entitled) is usually unclaimed,but barteredto a brotherin exchangefor a promiseof later security should they be widowed ordivorced.85 Fulfillment of dutiesto one'shusbandand largerfamily takeson a special importance,as doesproductionof sonsas a guaranteeforlater years.Thus,womenstrengthenwhateverlateral and vertical familyties they can, but in doing so they find themselvescarrying out rolesdictatedby the interestsof husband,sons,or husband'sfamily.

H husbandandwife haveseparatebut unspokenstrategiesto protecttheir interestsand security, so, too, does the family as a whole. Thedegreeof purdahobservedis part of family strategy.Purdahprovidesrespectabilityand status. It enhancesmarriageprospectsfor daughtersand helps maintain the family position and name-avaluableassetthatyields power and influence in dealingswith the community. Thus, al-though its demandsare costly and restrictive, when families becomebetteroff they tend to keepstricterpurdah.08

Marriage of sonsand daughtersis a highly plannedpart of family


strategy; as in most of the subcontinent,it is an arrangementbetweenfamilies. Early marriageenableslargerfamily intereststo dominatethoseof the young couple in the first 'few years. And where high fertility isvalued, it allows the family as a whole time to accumulatechildrendespitethe long lactationperiodsthat prevail in Bangladesh-astrategythat goesfar toward maximizing the chancethat children will not onlybebornbut also will survive pastthe high-mortality yearsof infancy.

. Different treatmentof sonsand daughtersalso reflects family inter-estsat work. Daughtersdo useful householdwork, but unlike sons arenot in a positionto bring in outsidewagesor providelater support.Undersuchcircumstances,it falls within family intereststo passon daughtersatan early agebut to hold sonsas long as possible-afactor that may con-,tribute to the largesex differencein ageat marriage.

In eachof thesecases,behaviorrational to the family helpsperpetu-ate a position of low autonomyfor women.

Fertility:ConstraintsandSupports

The high level of fertility in Bangladeshoccurs fairly uniformly acrossregionsand classes.Larger landownershaveslightly higher fertility thanaverage;poorerwomenandurban women tend to have fewer children,but all the differencesare small. Lower averagefertility doesseemtoprevail amongthe urbanmiddle class,but this group is of insignificantsize and is unlikely to absorblarge numbersin the next two decades.61

Although averagefertility varies little acrossgroups,within any groupof women variation can be large, reflecting mainly the biology of theprocess.For currently marriedwomenin their 40s in 1974, the distribu-tion by children everborn is shownin Figure4. About 50 percentof thewomenhadbornesix or more children, while 30 percenthadbornefouror fewer. Relatively small numbersof births in a completedfamily areby nomeansrare.

In the absenceof any obvious reasonsthat birth rates should bechanging,it hasbeenconventionalto assumea roughly constantlevel offertility over pastdecades(seeFigure 2). There are, however,year-to-year fluctuations-for example,a substantialdrop in the year followingthe1971war andagainafterthe1974famine.68

It is clearly importantto try to explainwhy fertility is at its currentlevel. Later in this sectionwe arguethat social and economicpressureson families combineto pushfertility upwardto a natural limit. Yet, witha female reproductivelife of 35 yearsor more and with early and uni-versalmarriage,why should that limit be a total fertility rate as low asseven?Ratesbetweeneight andtenhavebeenreliably recorded'in popu-lationswith laterageat marriagethanin Bangladesh.(Figure5 compares


the age patternof fertility in Bangladeshwith one such population, theHutterites of North America.) The answeris best given by listing themain factors, mostly biological and partly due to societalarrangements,that constrainBangladesh'sfertility below theseextremelevels.

Figure 4Distribution of Women Aged 40-49by Number of Children Ever Born,Bangladesh,1974

First amongfactorsdepressingnaturalfertility is the fact that womenin Bangladeshexperienceperiods of lactational amenorrhea(i.e., sup-pressionof ovulation as a consequenceof breastfeeding)that are amongthe longest ever recorded.Extendedbreastfeedingis the main reasonBangladeshis not reproducinglike the Hutterites. Lactational amenor-rhea accountsfor nearly 19 months of the average34-month intervalbetweenbirths. Sinceit is not difficult to visualizeforcesthatwould tendtq shortenbreastfeeding,thereis a constantthreathangingoverprospectsfor fertility reduction. H the lactational period were reducedso thatamenorrhealastedonly 11 months,the level prevailing in Punjab, totalfertility would be 25 percenthigher.60 (Of course,mortality would alsoincreasewith shortenedlactation becauseof the removal of the protec-tion that breastfeedingand birth spacing afford to infants. The net

w. Brian Arthur / GeoffreyMcNicoll

Figure5Age-SpecificFertility Ratesin Bangladesh,1974,Comparedwith anExtreme"Natural" Fertility Schedule

55

increasein surviving children could be quite small.7°) Nutritional inade-quaciessufferedby poorerwomen add slightly to the length of amenor-rhea.71

Less important as a depressantto fertility, but still significant, isabsenceof the husbandthrough death,divorce, or occupationalmigra-tion. We have already mentioned the high frequency of widowhoodtoward the end of the reproductiveyears. Husbandsare considerablyolder than wives, so that remarriageprospectsfor women are few (andare further reducedby social disapproval).A rough calculation using1961 marital statusdatasuggeststhat fertility could be 6 percenthigherif all husbandssurviveduntil their wives reachedthe end of their repro-ductive years.72 For divorce, a similar calculationgives a fertility effectof 5 percent.Physicalseparationof couplesasa resultof seasonalor otherkindsof occupationalmigration is difficult to measure,but not negligible.One study estimatesthat it lowers fertility by 10 percent.73 Differentialmortality andmigration probablyaccountfor most of the classdifferencein fertility.

As a factor in redUcingfertility, primary sterility appearsto be rela-


tively unimportantin Bangladesh.The 1961censusfound only 5 percentchildlessnessamongmarriedwomenover age45; the 1974RetrospectiveSurvey found 4 percent.74 Of course,female sterility would often be areasonfor divorce; so this categoryoverlapsthe previousone. It is notclearjust how muchsecondarysterility exists,but with thehighmorbiditylevels it may be substantial.7G This is probably the main reasonfor thequite large numbersof womenwith low completedfertility in Figure 4.

FiilalIy, contraceptiqnis not at presenta major factor depressingfer-tility in Bangladesh.The 1968-69National Impact Surveyrecordedonly3.7 percent current users of contraceptionamong married women ofreproductive age.78 U used effectively, contraceptionof this amountwould meana birth rate of at most 3 to 4 percentbelow its level in thetotal absenceof use. There is no reasonto supposemore than a slightincreasein theuserrateover the lastdecade.The problemis partly a lackof knowledgeof modemcontraceptionandvery limited accessto regularsuppliesandservices.The extent of demandfor contraceptionis a con-troversial issuestill to be fully resolved,but little evidencepoints to alatent demandof large proportions.77 An important experimentunder-takenin Matlab thanain ComiIla seeksto probe the limits of contracep-tive demandby a saturationdistribution scheme.Resultsindicatea ceil-ing of around17 to 18 percentof married women in reproductiveagesusing the profferedsupplies(a peakattainedafter threemonths),tailingoff after two yearsto less than 10 percentdespite continuedintensiveeffort put into the supply system.78 Abstinence, a sure contraceptivemethodavailableto all, doesnot seemto be muchusedas an intentionalmeasurefor controHing births. For example, the taboo against sexualrelationsduring breastfeeding,observedin many other cultures, is notstrongin Bangladesh,nor is tenninalabstinenceafter becominga grand-parent. (Breastfeedingitself doesnot seemto be usedconsciouslyfor itscontraceptivevalue,althoughevidenceon this point is weak. It is possiblethat the usualpracticeof a wife's going to live with her motherfor birthand subsequentrecuperationis sometimesextendedas a meansof birthspacing.)

While satisfactoryas an explanationof how completedtotal fertilitycomesto besevenandnot nineor ten, we havenot yet askedwhy fertilityratesshouldbe pushingagainstbiological limits despiterapidly worsen-ing man-landratios and increasingimpoverishment-trendsthat arewellrecognizedby the peoplethemselves.79 We would not of courseexpectto seerapid adjustmentto the new mortality conditions, but a genera-tion shouldbe a long enoughtime period to detectsubstantialsigns ofchange.Yet fertility continuesto remainhigh throughoutthe rural sector.

It is hard to believe that high fertility in Bangladeshis merely theresult of blindnessto choice. Certainly when we look at some of theindividual efforts to improve economic position by adopting whatever

W. Brian Arthur I GeoffreyMcNicoll 57

meanscanbe found, we seethat peasantsin Bangladeshasmuch aselse-where seem quick to senseopportunities for even marginal progress.Perhapsit could be arguedthat the lack of accessto modemcontracep-tion maintainsfertility at high levels. But we have also seenthat it ishardto give weight to this factor. Primitive meansof contraceptionhavealwaysbeenavailable,andmodemcontraceptivesofferedon a saturatiopbasishavegeneratedat besta lukewarmresponse.

It is at the level of economicand social pressuresthat fertility sup-ports are most likely to be found. It is difficult, however,to unravel thecomplexnetworkof incentives.Startingat the family level, it is likely thatin a situationcloseto subsistence,extra hands,especiallymale onesthatcanbring in a usefulwagebeforeageten, couldmakeasubstantialdiffer-enceto the family's well-being. Childrenare profitable as an investment.Evidencefrom the recentvillage study by Mead Cain (1977) supportsthis view: By age12 the averagemale child hasbecomea net producer;by age15 he has compensatedfor his cumulativeconsumption;and byage 22 he has repaid the investmentin himself and one sister. Parentsthereforerealize a net return on children, a return that would be lowerif rural wagesfell further, but one that would be higher if sisters'house-hold work were valuedand if later family supportwere included. Thisis not to saythat parentsperform any consciouscalculationson fertility,but an awarenessof the advantageof children as extra incomecan wellbe presentin high degree.

The finding that children are profitable after 15 or so years in asocietythat lives close to the margin may not be sufficient to show thathigh fertility makessense.The earlyyearsof childhoodarea consumptiondrain, andundercircumstanceswhereday-to.daysurvival dominatesanyfuture payoff, why should parentscarry the extra burdensof child con-sumption?It is here that we refer back to our discussionof the family.For the early years of marriage, we found that the patrilineal familylightens much of the burden of childbearing; later we saw that at noparticularpoint in the life cycle do incentivesagainstchildbearingbuildup. And women, in their special position of dependence,find sons thehighest form of security againstprivations of widowhood. Cain makesa further interesting point here. In rural Bangladesh,if a husbandisincapacitatedor diesbeforehis eldestson is old enoughto assumecontrolof family property, the family faces a period of extremevulnerabilitywhen assetsare easily lost through lack of male labor or through socialpredation.8o Futuresecuritythereforemakesit importantfor womennotonly to havesons,but alsoto havethem as early aspossiblein the maritallife cycle.

Economic interestsin producing children vary somewhatby class.For the relatively affiuent landowner,better survival prospectsof chil-dren are probablynot sufficient incentive to cut back fertility. Children

58 POPULATION AND ｄｅｖｅｌｏｐｍｅｾｔ IN BANGLADESH

can be educated,sentto cities, andusedto expandhis local power. Sub-division of land amongchildren poseslittle threat: Family holdings canbe augmentedthroughmarriageandby purchaseor foreclosureof mort-gages,and other occupationaloutlets are increasinglyavailable.

For those lower down, children provide, besidesdirect labor, theonly real securityagainstlossof holdingsandpenuryin later years.Thus,whetherparentsare concernedmostly for their own welfare or for thatof their childrenaswell, high fertility is an advantage.That the childrenthemselvesare then likely to inherit plots too small to supporta family, afactor assumedto favor low fertility among the peasantproprietorsofpreindustrialEurope, is necessarilyhere a secondaryconcern.The ex-tremeuncertaintiesof the local social andeconomicsettingdo nothing toencouragegrowth of peasantproprietorshipin Bangladesh.

At the bottom of society, lack of security overridesall other con-cerns.To the extentthat attachmentto a patronis possible,childrencanadd "pleading power" as claimants in the labor market.81 And at thislevel childhood (and maternal) mortality runs highest. Families countthemselveslucky if they canproduceenoughbirths so that evenone sonsurvivesto adulthood.82

It could well be arguedthat high-fertility behavior,thoughsensiblefor parents,runs counterto the interestsof the local society and, there-fore, that powerful socialpressuresmight ariseto opposeit. But for thisto happenthe local socioeconomicstructuremustbe able to transmitandsignal to parentsthe interestsof the communityat large. The patternofsocial organizationin Bangladeshthat we outlined earlier appearssing-ularly ill suitedto sucha task. Lines of social influencerun throughkinand patronagegroups,whosemembershipand territorial boundariesarefluid andwhosechief concernsare in servingthe economicand politicalinterestsof the dominantfamilies within them. High fertility poseslittlethreat to those interests. At a lower level the patriarchal family canenforceearly marriageand supporthigh fertility of the nuclearunit bysharingthe dependencyburdenof children and by awardingstatuswithachievementof motherhood.83

We have sketcheda picture in which the separateinterestsof ex-tendedfamilies, of most nuclear families, and of factional groupspressfertility to levels detrimentalto the communityat large. If eachof thesegroupsgainsseparatelyfrom its own actions,who then bearsthe costs?Threecategoriesof peopledo. The communityasdefinedby geographicallocation suffers through constantlowering of wages and use of scarceresources.Younger wives who are induced to deliver births at a ratebeyondtheir wishesand sometimestheir capacityalso suffer. But mostof all it is the next generationthat loses-thepeople who in 15 to 20years will somehowhave to find new meansof subsistence,move tomarginal land, or attach themselvesto the factional system.Fertility in


Bangladeshhas not fallen, becausetheseare precisely the three groupsthat have traditionally had no way to organizeto protect their interests:Unborn children by definition have no voice; women as yet have littlestandingoutside the family; and villages or other territorially boundedsocial units in Bangladeshhave a minimal role in the society.

Prospects

In describing the context of rural developmentand demographicbe-havior, we havebeenled to questionsof economicclass,social organiza-tion, family structure, and local government.\Ve have seen how theindividual aim of achieving security and a shareof the social productin a tighteningenvironmentcan act againstcollective demographicanddevelopmentinterests.And we havenoted the specialhistorical circum-stancesof the region that have left behind a power structurein whichnational ability to influence events quickly fades with administrativedistancefrom Dacca.

What does the future hold for Bangladesh,and what questionsofdevelopmentstrategydoesthe country face?We cannotexpectto arriveat a narrow rangeof likely outcomes,both becauseof the innate uncer-taintiesof the future andbecausemuch dependson the scopeand direc-tion of governmentaction. But we can layout the bounds of futuredevelopment,separatingthe plausible from the highly unlikely. Thisdone, we can trace some of the changesthat will affect the courseofdevelopmentwithin thesebounds.

The Broad Picture

Growth of population numbers must be expectedto show the usualdemographicinertia. If fertility and mortality change little from theirpresentlevels, the country would contain around160million peoplebythe year 2000; if therewere dramaticdeclinesin both (as set out in thetop panel of Table7), the country wouldstill have around 130 millionpeople.Catastrophesof the magnituderecently experienced,as Baldwin(1977) hasshown,leavethe courseof populationgrowth little altered.8t

Oneway or another,Bangladeshwill probablyhave to supporta popu-lation between60 and 80 percentlarger than its presentsize by the turn

of the century and one that may then (as now) still be increasingbymore than 2 million a year.

The main hopesfor supportingthesegreaternumberslie with rapidagricultural developmentthrough improvedwater control, high-yieldingcropvarieties,andothercomponentsof the GreenRevolutiontechnology.Earlier we notedthe apparenttechnologicalfeasibility of a three-or even

60

Table7PopulationProjections,Bangladesh,1978-2000

POPULAnON AND DEVELOPMENT ｉｾ BANGLADESH

Projection 1978 1988 2000

Rapidly declining mortality and fertilityBirth rate (per 1,(00) 45 31 20Death rate (per 1,(00) 17 12 8Population (millions) 85 109 128

World Bank "probable" projection aBirth rate (per 1,(00) 46 36 30Death rate (per 1,(00) 18 14 10Population (millions) 85 113 148

Continuing high mortality and fertilityBirth rate (per 1,(00) 47 43 40Death rate (per 1,(00) 19 18 17Population (millions) 85 116 159

a Bangladesh, Population and Nutrition Projects Department (1975, Annex 14-16); ad-justed for consistency with revised 1974 census total, and with 1974 Retrospective Surveymortality estimates.

fourfold increasein rice productionwithin a few decades;however,oursubsequentaccountof the rural settingshouldserveto temperexcessiveoptimism.It is highly likely that food productionwill lag far behindsuchpossibilities,althoughit shouldnot provetoo difficult to achieveincreasesthat more than keep abreastof populationgrowth. But for families inagriculturethe outlook dependsas much on distribution as on total out-put. Thosewith substantiallandwill profit mostfrom thenew technology,both by directproductivity gainsand,barringshifts in the presentpatternof local administration,by their influence over local irrigation and dis-tribution of neededinputs. Among marginal landholdersthe processofland impoverishmentwill continue,worseningpresentinequalities.Andfor the landless,the greaterlabor intensityof the high-yieldingcrop vari-etiesis unlikely to do muchto shoreup rural wagelevels faltering underpressureof a rapidly expandingwork force. Nonagriculturalemploymenton a scaleableto absorbsubstantialnumbersof thosepushedoff the landis difficult to visualize,but the paceat which such employmentcan becreatedwill largely determinethe welfare prospectsfor the bottom halfof therural population.

On the industrialsidewe sawa numberof opportunitiesfor sustainedif modestgrowth, diversifying from the presentjute manufacturingbase.Natural gaswill afford a largemeasureof energyself-sufficiency,at leastfor the next few decades,and alreadyservesas feedstockfor a growingfertilizer industry; however,other natural resourcesare lacking. Foreignexchangeshortagesmay be easedsomewhat by new revenuesfrom


natural gas exports and by a lesseningneedfor food imports. (Directforeign assistancenet of debt ｳ･ｲｶｩｾ･ will probably not again reach thelevels of the early years of independence.)Whether Bangladeshcanemulatethe few developingcountriesthat have mc..vedsuccessfullyintomanufacturedexports is doubtful. Uncertaintieshere, besidesdomesticconditions, include future economicand political relations with India-with its enormousmarket and strong industrial base-andthe scopeoftrade concessionsin any new economic order beiween rich and poornations.Themost optimistic projectionof growth in Bangladesh'sindus-trial economyis probablythat of FaalandandParkinson(1976, p. 174),who think that an annualgrowth rate of 10 percentshouldbe attainableunderfavorableconditions-arate that would in 25 yearscreatea sectorcomparablein size to agriculture.A more plausibleprojectedrate of in-creasemight be thatof Indian industryoverrecentdecades,an averageofabout6 percentperyear.85 As yet thereis little signof rapidgrowthgettingunderway: Presentindustry,operatingin a climate of regulatoryuncer-tainty, price changes,and intermittentshortagesof foreign exchangeandspare parts, has barely recoveredfrom the 1971 war.86 Moreover, asimportantas aggregategrowth in looking to Bangladesh'sindustrial de-velopmentare choicesconcerningscaleof enterprise,capital intensity,public or private ownership,and decentralization.Continuing politicaluncertaintiesleavethesechoicesopen.

What of the urban future of Bangladesh?If cities continue to in-creaseat their recentrate,by the year2000about30 percentof the popu-lation would live in urbanareas.Even at this comparativelymodestlevelof urbanization,cities would have to absorbanother35 million people.Put anotherway, the economywould haveto supportanother15 to 20cities the presentsize of Dacca-or,more likely, a few vast urban ag-glomerationssurroundingDacca-Narayanganj,Chittagong,and perhapsKhulna. Under realistic assumptionson economic trends, this rate ofurbanizationis likely to bean upperlimit. But evenwith slowereconomicexpansion,cities may still grow at close to this pace; for whatevertheincreasein their industrial, commercial,and governmentallabor force,they will inevitably also remaina dumpinggroundfor the extrudedruralpoor. Urban growth dependson more than the speedand natureof in-dustrialization.

In sum, allowing for a certain measureof speculation,the presentsituation can continue. In all probability, by the year 2000 the countrywill remainheavily rural; population..vill be largerby abouttwo-thirds;agriculturaloutputperheadmay havegainedmodestground;and a highproportionof therural population,probablymorethan half, will be land-less. Per capita income will be higher, markedly so at the top of thedistribution. For the majority of families, life will go on much as before,perceptiblybetterif things go well, harsherif they do not.


We haveput roughboundson the pathsof severaldevelopmentvari-ables: population,agricultural output, industry, and urbanization.Moreimportantthan the actual levels attainedat the end of the century,how-ever, are the quality and direction of the developmentprocessat thattime and the valuesandorganizationof societythat supportthis process.These deeper aspectsof the economic and demographicsystem willdeterminewhetherthe outlook in the twenty-first centurywill be one ofpromiseor of disaster.In the yearsjust ahead,much dependson the kindof social changetaking place in the countryside,on the possibilities ofrapid fertility decline,on the effectivenessof governmentinterventionindevelopment,andon whetherthe local economicsystemcanbe organizedfor sustainedgrowth. To thesequestionswe now tUl11.

Social Change

We have alreadypointed to the emergenceof a more pronouncedclassstructurein rural Bangladeshas subsistencepeasantsare shakenout oftheir place in the economyand addedto the landlessor near landless.The new agricultural technologywill in all likelihood hastenthis process,both by directly increasinginequality and by giving further impetus tothe commercializationof farming-sharpeningthe once hazy line be-tweenlandownerandwagelaborer.

It is difficult to predict the outcomeof increasingclassprominence.Certainlyit will do much to weakenthe socialstability resultingfrom kinand patronageties. But factionalism is unlikely to disappear.Even in acommercialeconomy,control overemploymentandmuchelsewill remainin the handsof a few-a basisfor continuedrecruitmentof dependents.Rural classconflict will probablybemirroredin a strengtheningof antag-onismsbetweennational political groups, and the possibility that thisconflict will lead to a radical reorganizationof the entire society cannotberuledout.

New avenuesof social mobility arebeginningto appear.While themost deprived elementsof the rural population have little chance ofimproving their lot, even if they move to the towns,for others there areincreasingopportunitiesfor upwardmobility that bypassthe traditionalstatushierarchyof rural society. Education is one route;attachmenttothe lower fringes of the administrativesystemis another.The new roleof 'orokering"betweengovernmentprogramsand the local populationisparticularlypowerful, althoughfor the mostpartsuchopportunitieshavesofar fallen to the traditionalvillage elite.81

Thesecontactswith the more modem parts of the economyserveto channelinto the countrysidenew attitudestoward such things as therole of women, the value of education,age at marriage,and contracep-tion. Governmentextensionworkersconcernedwith particularprograms(irrigation, family planning,and so on) are anotherconduit into village


life; and thereis a small but increasingamountof travel betweenvillageand town.88 But the tangible realities of life may limit the degreetowhich new ideas can take hold. For the considerablefraction of thepopulation at bare subsistencelevel, the notion of modernity scarcelyapplies.For them,a mode!-'I1izingsocietymay well be one in which theirformer place in the schemeof things no longer exists.

Deepeningclassdivisions,as well as the slow appearanceof modernattitudes,work to changethe position ofwomen.\Vomenwho are pusheddownward in society through loss of husbandor land join those whocannot afford the costs and restrictions of purdah and become freeto take on new roles. Women'scooperatives,for example, have foundtheir recruits typically to be drawn not from the elite but from thewidowed, the landless,and the destitute.89 But there appearslittle pros-pectunderpresentarrangementsthat changein women'sroleswill "trickleup" from the bottom of society.Gradualdiffusion of new attitudesfromthe small, educatedelite is the more likely sourceof change.

Under this pattern of change,and assumingno abrupt moves insocialorganization,it is difficult to seewherestrong incentivesto restrictfertility will comefrom. For most families, children would continue forsometime to provideincomeandsecurityin a shakyenvironment.Largeimprovementsin accessto family planningserviceswould of coursehelp:Mopping up the residueof existing or easily drummed-updemandforbirth control in Bangladeshmight lower fertility by about15 percent-toa birth rateof around40 per thousand.Increasesin ageat marriagecouldalso havea significant impact in reducingfertility. The changeswe havedescribedposelittle immediatedisincentiveto early marriage,althoughtheir effectsmay mount up. To offset such factors,however,is the possi-bility of shorteningperiodsof lactation.

Fertility declines are not of value irrespective of how they areachieved.A fall in the birth rate, for example,would be one likely con-sequenceof a further deterioration of lUral conditions-a product ofworseninghealth status,physical separationof couplescausedby labormigration, andgeneraldisintegrationof stablefamily life. Children couldevenceaseto be a paying propositionundersuch circumstances,addingvoluntarypressureto this involuntarydecline.po

To summarize,we would not look for a major decline in fertility inBangladeshin the next 10 to 15 years. A more likely pattern would besomewhatlower fertility for that part of the population whosecircum-stancesareworsening,togetherwith a slow chippingawayof the fertilityrate of the rest as contraceptivedelivery spreads,women define newroles for themselves,and mortality, especiallyinfant mortality, improves.This not very hopeful outlook supposescontinuationof the presentsocialandeconomicarrangementsin rural areas.Under different circumstancesdemographicprospectscould be radically altered.


Government Action

We have explored the likely consequencesof technologicaland socialchange;what of developmentthrough direct governmentaction? Thecharacteristicapproachto developmentpolicy of the Bangladeshgov-ernment,as of most others,could be describedas programmatic.Succes-sive governmentsover the past30 yearshavesetup a variety of programsdesignedto promoterural developmentand restrainfertility. The Inte-grated Rural DevelopmentProgramme,growing out of the Comillacooperatives,and the Rural Public Works Programmewere mentionedearlier. Othersinclude the Village Agricultural and Industrial Develop-mentProgramme,the ThanaIrrigation Programme,and a wide array ofpilot programsconcernedwith agricultural extension, literacy, health,andso on. In family planning,a large-scaleoperationhasbeenmounted,initially involving clinic-basedservicesandlaterfemaleparamedicsbasedin thevillages.01

The record of theseprogramsis at best describedas mixed. Manyhaveshownencouragingresultson a small scale,when closemonitoringand supervisionwere feasible. But as the programshave beenextendedgeographically and to encompassmore activities ("integrated"), theresultshavebecomelessimpressive.For the most part, as we haveseen,the programshave comeup againstlocal realities-arural social systemthat worksafter a fashion,possessesa measureof stability, and is able toadaptto its own purposesthe best intentionsof planners.

There are also instancesof privately sponsoredrural developmentefforts in small regionsthat have been reported to be very successful.Theseinclude producercooperatives,health clinics, women'sorganiza-tions, and even village-level efforts to limit populationgrowth.92 Whileof considerableinterestin themselvesandfor their potentialto shedlighton feasiblemechanismsby which communitiescan be organized,theseefforts haveyet to demonstrateanyvalueaspilot projectsfor countrywideprograms. The effectivenessof the Swanirvar (village self-reliance)movement,recently initiated by the government,may providean inter-estingtest of this question.As this movementsolidifies into a program,there seemfew reasonsto expect it to escapethe problems that havebesetits predecessors.But its promiseif it cando so is considerable.93

In spiteof their mixed success,mostgovernmentprogramsdo pushdevelopmentin the right direction. Intensified and well-managedcam-paignscan provide somerelief for rural underemployment,disseminatenew agricultural techniqueswith greaterrapidity, and reduce fertilityslightly. And program activities themselvesdo eventually affect socialstructure.Thereis no doubt, for example,that the manifold agriculturaldevelopmentopportunitiessketchedout by NoazeshAhmed (1976) ifacted upon would, through the resulting economic change, radicallyalter rural social organizationand probably also demographicbehavior.

W. Brion Arthur / GeoffreyMcNicoU 65

But thereremainsthequestionof whethera programmaticapproachis appropriateto Bangladesh's developmentpredicament.Plannersnatu-rally tendto think of policy interventionsin termsof inputs andoutputs.For largedevelopmentprojects-Hoodcontrol works, irrigation channels,industrial plants-thisis the obvious way to assessfeasibility and effec-tiveness.But with lesscentralizedandlessmaterialactivities-in particu-lar, rural developmentand population control-maintaininga similarapproachmay distort emphasisand obscure important problems.Wehaveargued that thecritical policy issueslie in the areaof local politicsandsocialorganization:how the interestsof local factionsand dominantfamilies are pursuedat the expenseof the community as a whole, andhow new political and material resourcesenteringthe rural societycanserveto strengthenratherthanerodethis pattern.Thesestructuralissuesarenot merely "obstacles"to programdesignand operation,to be takencareof, as it were, at the stageof implementation;rather, they are thestuff of developmentitself andshouldbethe main objectsof policy.

StrategyandStyle

In speculatingon future prospectsand government'spart in them, wehavedeliberatelyavoidedproposingany packageof nationalpolicies. Ina country whosenational governmenthas at best only limited purchaseon local be:lavior,mostpolicy proposalswould necessarilybe moot. Sug-gestions,say, for strengtheningcooperativesor for raising the legal mar-riage age would chieHy serve to exposetheselimitations. Bangladesh'sexperiencewith cooperativesis rivaledby thatof few othercountries,andcontinuedfailings in this areaare not for want of understanding.Andfor marriagethe legal minimum agefor' womenalreadystandsat 16, inspite of which most are married at younger ages.94 In these finalremarks,then, we confine ourselvesto somegeneralobservationsat thelevel of developmentstrategy.

We would concludefrom our analysisof Bangladesh'sdevelopmentand demographicsituation that the country's presentproblemsare lessthe result of not having some "right" combinationof national policiesthan they are the consequenceof a local societalstructurethat the gov-ernmentcannotquitereach.At its lowestlevelsthe presentadministrativesystem mirrors the interests of the larger landowners and factionalleaders.Thus, national governmentintentions and local economic anddemographicrealities typically work at crosspurposes,impeding devel-opmentresults.An effective local administrativesystemable to harnesspowedul but presentlyconflicting social forces could do much to alignindividual behaviorwith the collective interest.

There are many ways in which effective local organizationmightbe achieved,and severalAsian modelsmight be adaptedto the specialneedsof Bangladesh.The recentexperienceof Indonesiaunderthe "New


Order" regimeindicatesoneoption: pressinggovernmentinfluencedown-ward throughsocietyand designinglocal incentivesto ensurethat com-munity interests(broadlydefined,it is hoped)dominate.In Java,equalinarea,population,and straitenedcircumstancesto Bangladesh,incentiveswere imposedthrougha greatly strengthenedadministrativesystem;thishasaheadyyieldedsigni6cantresultsin birth control,but it is still too earlyto assessits economicoutcomeor to weigh the long-runimplicationsof thepervasivemilitarization of local governmentthat underliesit. An alterna-tive strategy,lessharshin somerespects,harsherin others,is to limit admin-istration to levels above the local community, making small territorialunits responsiblefor their activities and relying on internal social pres-suresto bring individual behavior in line ｷｾｴｨ the community ｩｮｴ･ｲ･ｳｾＮ

This strategy,similar to the one that operatesin rural China below thecommunelevel, would presentproblemsin Bangledesh,where naturalterritorial units are weakandwherecurrentsocial changeis, if anything,further eroding what territorial cohesionthere is. Yet possibilities exist.The Chinesecommune,a purely administrativeunit, is roughly similarin sizeto the Bangladeshunion. The commune'sconstituentvillages (pro-ductionbrigades)andwork groups(productionteams)haveat bestonlyweak counterpartsin Bangladeshat present,but it might prove possibleto make either gram or para into solidary social units in a Bangladeshrural developmentstrategy.911Rural developmentmodelsthat involve lessall-embracingorganizationalforms are also available. In the Bangladeshsettingan option may be to build separateorganizationalbasesto pro-mote rapid growth of agricultural output and to ensurethe rural land-lessof accessto theeconomy.

Whatevermeansis chosento reachdown to the local level, effectivesocial changeundoubtedlyrequires a local setting much better able toafford securityto the individual. A major themeof this paperhas beenthat the interestsof rural peoplein an uncertainand unhealthyenviron-ment are best servedby behavior that guaranteessome accessto localproduct and that lowers risks. Attachment to patrons, exploitation ofscarcelocal resources,adoption of roles of dependency,maintenanceoffamily position through suchmeansas purdah,and use of high fertilityto insureagainstloss of land, children, or husbandmay yield accesstothe local distributive system; but above all they provide security. Inaggregate,though,thesemodesof behavioract againstcollective interestsand lend power and stability to the forces behind the presentpredica-ment. Much could be done to provide security at the local level-eco-nomically, through surer accessto employment,credit, agricultural in-puts, and irrigation; and demographically,through measuresthat pro-long the lives of husbandandchildrenandreducethe gapin agebetweenhusbandand wife. But more important than specific measures,if thenew local systemis to help break down the uncertaintiesthat preserve

W. Brion Arthur / GeoffreyMcNicoU 67

high fertility and stifle individual initiative, it must provide a setting inwhich the individual looks to the communityand his position within itfor supportand security,and lessto the capriciousfavors of landlord orpatriarch.

This needfor effective structuralorganizationhas its counterpartinthe industrial sector.Effective industrial developmentwould call for asecureinstitutionalenvironmentfor production,a needthat goesmuchdeeperthan the usualdemandfor infrastructure.

Thetypeof "intennediatelevel" strategy-directedat a level betweenthe nationalgovernmentand the individual-thatwe seeas essentialtotackle the problemsof Bangladeshis more a matterof structuralchangethan of programmaticpolicy. Even with abrupt transfers of politicalpower, it would be no short-tenntask. What is missing at presentinBangladeshthatwould greatlyeasesucha taskis a coherentdevelopmentstyle-anoverall strategythat works toward a singlepurpose,that alignsinterestsin the samedirection, and that cuts through conflicting objec-tives. Thepresentarrayof looselyorganizedprogramsandad hoc incen-tives is morea complexreactionto thepressuresof urbaninterestgroups,internationalagencies,andthe rural powerstructurethana singlemindedinstrumentof development.

Thenext yearsin Bangladeshcouldseethe emergenceof very differ-ent types of political regimes.Whateverthe political color of the gov-ernment,its successin developmenttennswill likely be measuredby itswill and administrativereach to cut acrosslocal interestsand influenceeconomicand demographicbehaviorfor the national good. If this doesnot comeabout,the presentsituationcancontinuefor a while andpopu-lation growth with it; but the possibility of real developmentwill recedefurther into thefuture.

Notes

The authors wish to acknowledge thehelpful commentsof many colleaguesonearlier drafts of this paper. In particularwe would thank Monowar Hossain,AbuAbdullah, Shapan Adnan, MohiuddinAlamgir, Mead Cain, Lincoln Chen, andother presentor fonner staff membersofthe BangladeshInstitute of DevelopmentStudies; and John Bongaarts,AdrienneGermain, ｾｯｳｨ｡ｲ｡ｦｦ Hossain, Moni Nag,and Veena Thadani.We expectnone ofthese colleagues to agree with all wehave said, and some to disagreewithmuch. A preliminaryversionof this paperwas discussedat the Workshopon Coun-try Reports on Populationand Develop-

ment, held in Bergen, Norway, Septem-ber 1977.

1. Moghul influence dates bacle tothe thirteenth century. On the economyof pre-Muslim Bengal see Majumdar(1943), chapter 16. Later accountsarecited in N. Ahmad (1968), chapters:h5.

2. Contributingto the disappearanceof East Bengal's textile industry werethe effects of the Great Bengal Famine(1769-76), followed by severalyears ofdevastating ftoods and famine in the17805, the subsequentreorganizationofthe revenuecollection system under thePermanent Settlement policy, and the


high tariff imposed to protect Britishproducts. See N. Ahmad (1968), pp.100-104, and M. A. Islam (1968). Asmall jute hand loom industry grew upfor some decadesin the nineteenthcen-tury, only to be supplantedby the Cal-cuttamills.

3. Historical data on city size areoften suspect,but there is evidenceof asubstantialpopulation decline in Dacca.Estimatesare200.000in 1780 and about70,000 in 1838. See N. Ahmad (1968),p.104.

4. A UN estimateof material dam-age causedby the war was US$1.2 bil-lion; lossesin the transportsectorand inhousing were especially severe. SeeFaaland and Parkinson (1976), p. 12,and N. Islam (1974). Other disruptionsresulted from the departure of WestPakistanimanagersand civil servants.

5. For two accountsof policy dUringthis period, before the denoumentbe-cameapparent,seePapanek(1967) andHaq (1963). Estimatesof the transferofresourcesfrom East to West Pakistan(highly dependenton exchangerate as-sumptions) range from US$1.5 to 3.0billion for the period 1947-69 (Faalandand Parkinson,1976, pp. 7-8).

6. Historical productivity dataareofcourse scarce, although indications canbe drawn from relative constancyof cul-tivating techniques. Some studies ofyields of the major crops from 1892 to1947 are noted by N. Ahmad (1968), p.191, who concludes:"Therefore,at best,productivity and yields may be regardedas constant,though betrayinga tendencytowards deterioration."Rice yields haverisen only slightly in subsequentdecades-about 15 percent between 1952 and1974 (N. Ahmed, 1976, p. 20). Juteyields show no clear trend over the past30 years.

7. PermanentSettlementgave rev-enue collectors proprietory rights overlarge estates (the zamindari system).SeeTepper(1970),chapter1, and M. A.Islam (1968) on the background andeffects of this policy. It is of courseeasy

to exaggeratethe impact of the colonialintervention: The likely bias in this di-rection in both domestic and Westernscholarshipis noted by Morris and Stein(1961).

8. In the late nineteenthcentury anetwork of local councils (some nomi-nally elected) was set up at the districtlevel and below, but turned out to berelatively powerless.The lowest level ofcouncil covered an area of several rev-enuevillages, calleda union. In 1919 theVillage Self-GovernmentAct establishedelected Union Boards with limited au-thority over local administration. SeeTepper (1970) for a useful institutionalhistory of preindependenceBangladesh.

9. See N. Ahmad (1968) and er-Rashid (1967) for comprehensiveac-counts of the physical geography ofBangladesh.

10. Of total croppedarea about 80percentis rice, 4-5 percentjute, and 2.5percent pulses. No other crop occupiesmore than2 percentof the area (Bangla-desh,Bureauof Statistics,1977).

11. The Ganges-KobadakIrrigationProject in Kushtia and Jessoredistricts,underconstructionsincethe 195Os,is themost notable. A major problem withlarge-scale water control is that thetopography dictates that necessarystor-age dams and canals would have to belocatedoutsideBangladesh,in India andNepal.

12. The FarakkaBarrage,completedin 1975, diverts water from the Gangesthrough its Indian distributary, theHooghly, to Hush the port of Calcutta.How much water will be diverted andwhat effect this will haveon Bangladeshwas an important point of contentionbetween Bangladesh and India beforean agreementwas signed in 1977. Thedispute was referred to the United Na-tions in 1976: Diversionof Gangeswaterin the dry seasonof that year, accordingto the Bangladeshgovernment,"had un-precedentedecologicalandeconomiccon-sequences.... Irrigation facilities, partic-ularly by pumps and shallow and evendeeptubewells,havebeenhamperedand

W. BriaB Arthur / GeoffreyMcNicoU

pumps rendered inoperative. Some400,000 acres of agricultural land havebeenaffected....River navigationwasdis-rupted. Fishery,forestry and ｬｩｶ･ｾｴｯ｣ｫ re-sourceshavebeenseverelydepleted.Thelevel of salinity has increased... serious-ly jeopardizing power and agriculturalproduction." (United Nations,1976)

13. Generalaccountsof the demog-raphy of Bangladeshinclude A. Ahmed( 1962) and M. R. Khan and A. R. Khan( 1975). The earlier classic study of thesubcontinentby Davis (1951) is charac-terizedby a lack of detail on EastBengal.

14. From 1872 to 1921 populationgrew at an averageyearly rate of 0.76percent,mostly through natural increase.For comparison,the Indian growth ratefor 1871-1921was0.36 percentperyear.The higher rate recorded in Bangladeshmay reJlect censuserrors or more rapidnatural increase;net migration into theregion seemsto have beenneligible. SeeDavis (1951), pp. 27, 109-111,and, onmigrationspecifically,M. R. Khan (1972b).

15. Death ratesmay havepeakedto50, 50, or more per thousandin times ofseverefamine or epidemic (Davis, 1951,chapter5). The officially recordeddeathrate for 1918, during the influenza epi-demic, was 63 per thousand (Davis, p.33). The most careful recent estimate(19.4, for 1961-74)comesfrom the 1974RetrospectiveSurvey (UK Ministry ofOverseasDevelopment,1977).Theofficialdeathrateseriesis given in the StatisticalPocketBook at Bang14desh1978.

16. Improvementsin transport andcommunication in themselves probablydid asmuch to increaseexportsandfacili-tate speculation in food. supplies as toevenout price levelsthroughinterregionaltrade. From the late nineteenthcentury,however,.governmeDtlaissezfaireattitudestoward famine were slowly replaced bya more activist policy (relief works, sus-pensionof land revenue,and so on), sothat as Bhatia (1967), p. 270, puts it,Faminewas transformedinto Food. Prob-lem.

17. Robinson (1967), pp. 14-15,and Davis (1951),chapter5.

69

18. Birth ratesin the rangeof 45 to50 per thousandaregiven by the 1962-65PopulationGrowth Estimationproject,byvariousotherdemographicsurveysin the1960s, and most recently by the 1974RetrospectiveSurvey,which estimated48per thousand.SeeM. R. Khan and A. R.Khan (1975) and UK Ministry of Over-seasDevelopment(1977). An estimateof42 is made by Sirageldin, Norris, andAhmad (1975), based on the 1968-69National ImpactSurvey. (The discrepan-ciesare not a productof shifts in agedis-tributions since they also show up inestimatesof total fertility.) See Potter(1977) and Blacker (1977) for discus-sionsof someof theerrorsin thesevariousresults.Note that if, asseemslikely, fertil-ity was low and infant mortality high inthe 1940s, birth rates may well havedroppedbelow their long-run averageinthe late 1960sas the small 1940scohortpassed through its peak reproductiveyears.

19. In Partition it is estimatedthat2.6 million Hindus emigrated from theregionwhile 700,000Muslimsimmigrated.During the 1971 war some7 to 10 millionpeople (chiefly Hindus-probably overhalf of Bangladesh'sHindu population)fled to India,with perhapsa million ofthem remainingin India after peacewasrestored.There was also a much smallerexodusof Hindus at the time of the 1965Indo-Pakistan War. See M. R. Khan(1972a,1972b).Khan speculatesthat netmigration over the 1961-72period couldhave amountedto more than 2 million.But even that numberis only a tenth ofthe actual population growth in Bangla-deshin that period.

20. SeeN. Ahmad (1968),pp. 300-303.

21. Caution is neededin tracing therecent population growth differentials tonew activity in agriculture.The data arenot fully reliable (the 1974 census inparticular was taken under difficult cir-cumstances), and the convulsive migra-tions of the early 1970sprecludeany sim-ple interpretationof growth rates.

22. On embankmentsand their im-


pact and on the forest ecology, see N.Ahmad (1968),p. 37 andchapter7. Eco-logical problemscreatedby the new ricetechnologyinclude increasedrisk of cropdiseaseand pest damage with varietalspecialization, eutrofication from heavyfertilizer use, harmful residuesof pesti-cides and weedicides,and the impact ofdeeptubewellson the water table.

23. See Faaland and Parkinson(1976), chapter2. The Bureauof Statis-tics index of real wages(1969-70= 100)fell steadilyfrom 108 in the mid-1960sto65 in the early 1970s (and to 49 in thefamine year 1974-75);in 1976-77it hadrisen to 71 (Statistical Pocket Book ofBangladesh1978).

24. Gustav Ranis (1974), p. 855, ismore hopeful than we on the prospectsfor export processingzones, leading inturn to decentralizedlabor-intensivein-dustrial production for export: "It is byno meansunrealisticto expectBangladeshto ultimately becomean important com-petitor in this area, a field vacated byTaiwan, Korea, Singapore,Hong Kong,etc. whose wagesare rising as a conse-quence ofthe exhaustionof the labor sur-plus."

25. The World Bank report is un-published; its findings are referred to byFaalandand Parkinson(1976), p. 132.

26. Area under high-yielding varie-ties had reachedonly 15 percentof ricecropped area by 1975-76 (Bangladesh,Bureau of Statistics, 1977). Inputs haverecordedrapid increases,but from a verysmall base.Datafor the early 1970sshowmore than 30,000 low-lift pumps in use,a tenfold increaseover a decade,and acumulative total of some5,000 tubewells(N. Ahmed, 1976,p. 25). Eachpumporwell on averagemight irrigate 15 to 20hectares.About 6.5 percentof cultivatedarea is now irrigated by low-lift pumps,1.2 percentby tubewells.Altogether,totalcroppedareaasa percentageof cultivatedarea has grown from 127 percent in theearly 1960sto nearly ISO percentat pres-ent. Fertilizer consumption (in terms ofnutrientvalue) quadrupledin the decade

1965-74 (Bangladesh,Bureau of Statis-tics, 1977).

27. Access to land is defined morepreciselyby effectivelandholding-thatis,land owned and cultivated, adjustedforthe amountsmortgagedandsharecropped(in or out). See Wood (1976), p. 69.Noncultivator occupationsare still onlya small fraction of the rural labor force.Fishing and livestock husbandryare typ-ically secondaryoccupationswithin farmor agricultural laborer families. Nonagri-cultural occupations (chiefly commerceand services)made up 17 percentof therural labor force in 1964,accordingto theMasterSurveyof Agriculture of that year.

28. Data on the agricultural laborforce and distribution of holdings aretakenfrom Alamgir (1977),Table 6, andthe 1968 Master Survey of Agriculture.The categorizationof farm families is ofcoursehighly schematic,ignoring the im-portantconsiderationsof family size, life-cycle stage, local agricultural conditions,and nonagriculturalincome, all of whichwould influencethe economicsignificanceof a particularsize of landholding.Manyvillages in the core areasof the countryhave no holdings above 3 hectares;forlocal purposesthe "surplus" cutoff shouldbe lower.

29. Cultivated areagrew by SO per-cent in the first four decadesof this cen-tury, but has changed little since the1940s. Average size of holding is con-sideredto havediminishedover the sameearlierperiod (N. Ahmad, 1968,pp. 184-187,195).Cain (1977) estimatesa sub-sistenceholding at 0.65 hectaresfor anaveragehouseholdin an averageyear.

30. In fact, Table 2 suggeststhatthere has been some squeezingout ofsmall farms. There is also evidenceof arising proportion of land operated bytenants-from18 percentof farm areain1960 to 25 percent in 1974 (Alamgir,1975, p. 269). On the actual proceduresof land transfersand their frequencybysizeof holding,seeAlamgir ( 1977),chap-ter 4. Bertocci (1970), chapter 4, dis-cussesthe economicsof land mortgagingand foreclosure. Mead Cain (personal

W. Brian Arthur / GeoffreyMcNicoll

communication)notesthat the processofupward economicmobility, althoughlesscommon than its opposite, does occur:Even originally landless agriculturalworkers,by effort and luck, cansometimesascendthe ladder.

31. The available data on landless-nessdo not constitutea consistentseries.The 1964 Master Survey of Agricultureestimatedthe proportionat 14 percentofthe agricultural labor force, whereastheSIDA/ILO Reporta decadelater recordedmore thandoublethis proportion (Abdul-lah, Hossain, and Nations, 1974). The1951 and 1961 censusestimateswere 14percentand 17 percent(Censusof Pakis-tan 1961, vol. 2); the 1977 Land Occu-pancySurveygives27 percent(StatisticalPocketBookof Bangladesh1978,p. 121).

32. Basedon the censuslabor forcedataand a smoothedseriesof proportionslandless. For the male labor force con-sideredalone,the 1961-74growth wouldhave beenabout 4 million, with 3 to 3.5million taken up by increasinglandless-ness.

33. Findings from the Dacca Uni-versityVillage StudyGroup (Adnanet al.,1977).

34. In addition they may be able toplace one or more members in securenonagricultural occupations, as clerks,teachers, and so on. See M. A. Huq( 1976),chapters4 and 7, for a discussionof classdifferencesin occupationaldiver-sification amongpeasantfamilies.

35. Cain (1977), pp. 206-207, de-scribesthe advantageto landholdingfam-ilies of providing most or all of theircultivating labor themselvesrather thanhiring workers: Family labor reducestheoften critical cash-Bowproblems,especi-ally at harvest time, and does nat needthe samesupervision.

36. The growth of migrant agricul-tural labor in Bangladeshis a particularlyimportant phenomenon.On its effect onthe local labor marketseeWood (1976),pp. 152-156.

37. The information on social orga-nization presentedhere andassembledin

71

Table 3 is an amalgam derived chieByfrom A. Islam (1974), Bertocci (1970),and the working papers of the DaccaUniversity Village Study Group.

38. Normatively, a bari would com-ｰｲｩｳｾ a man, his wife and sons,the wivesof marriedsons,grandchildren,unmarrieddaughters,and perhapswidowed sisters.A bari's land, however,may alreadyhavebeendivided amongits constituenthouse-holds even before the deathof the head,and conBict amongbari members,some·times leading to litigation, is fairlycommon.

39. See,for example,Dacca Univer-sity Village Study Group (1975).

40. The gramor '10calvillage," rele-vant in discussing social organization,doesnot alwayscorrespondto the mau%Q,a revenueunit also called a village thatwas set up in the nineteenthcentury, orto the more recently defined "censusvil-lage" created for statistical purposes.Bertocci (1976), p. 161, cites 1963 datafor the Kotwali thana (see note 47) ofComilla district showing463 gram (aver-aging 340 persons).whereasthere wereabout200 mauzaand 250 censusvillages.Investigationsprecedingthe 1975 WorldFertility Survey,however,found frequentcoincidence between the administrativeand socially acceptedunits.

41. A. Islam (1974), p. 75. The re-sidual identification, however, can beimportanton occasions:ConBictsor feudsbetweenvillages are not unknown.

42. See N. Ahmad (1956) for a re-gion-by-region description of settlementpatternsin Bangladesh.Nucleatedvillagesare the rule in the moribund delta areas.

43. The diminishing importance ofkinship groupsandshamaiis discussedbyAbdullah, Hossain,and Nations (1976),p. 217; Tepper (1976), p. 36; and Wood(1976), pp. 147-149. On shifts in agri-cultural arrangementswith commerciali-zation seeAlamgir (1975).

44. Wood (1976), pp. 154-1M.Many suchpeoplealso end up amongtheurban homeless:See Farouk (1976) fora descriptionof the processof extrusion


from the countrysidebasedon individualcasehistories.

45. An elaborate investigation intorural socioeconomicstructuresandchangein Bangladesh,now unde" way at thf'Bangladesh Institute of DevelopmentStudies,should substantiallyfill out ourinformation in this important area.

46. See in particular Adnan (1976)and Dacca University Village StudyGroup (1975).

47. Unions-have an averagearea ofabout 35 km2• The next higher adminis-trative unit, the thana, covers about 10unions, with an averagepopulation of160,000and averageareaof 300 km2•

48. Sobhan(1968),pp. 77-88;Tep-per (1970),p. 217; and A. Islam (1974).Rahman (1962) gives a systematicac-count of the Basic Democraciesschemeof 1959, which set up the union councilsand similar committeesat the thana andhigher administrativelevels. Union coun-cil members representelectoral wards,which do not necessarilycoincide withany social unit-sometimesspanningsev-eral villages, in other cases intersectingthem.

49. "[The well-off farmer1caninvesthis surplus money without any risk orlabor by just loaning it to his neighborswho are small farmers.Each year he getsalmostcent per cent return and his prin-cipal still remains unpaid. If he savessome paddy, he can loan it from harvestto harvest and get back one and a halfmaundsor morefor onemaund.So, whenthis surplusfarmer is told that he shouldusebettermethods,betterseedsor betterfertilizers for more production from hisland, he is not interested."(A. Abdullah,1976, p. 89, quoting a Comilla report of1963)

SO. The"Camilla experiment"and itsvarious progenyhave an extensivelitera-ture. For assessmentsfrom variousstand-points see A. H. Khan (1971), Raper(1970), and Abdullah, Hossain,and Na-tions (1974).

51. On the Rural Public Works Pro-

grammeseeThomas(1968) and Sobhan( 1968).

52. Abdullah, Hossain, and Nations(1976) assess these possible conse-ｾｵ･ｮ｣･ｳＮ

53. Data from 1974 census. SeeMiranda (1977). Two other urban areasexceeded 100,000 population: Mymen-singh and Rajshahi.

54. Alamgir (1973),p. 28.

55. On maniageagesseeCensusofPakistan1961,vol. 2, p. 111-10and Elahi(1974),p. 5. The 1968-69ImpactSurveyestimatedthe median age at marriageofa sampleof marriedwomen underage55to be 13.2 years,with three-quartersmar-ried beforeturning 14 andonly 5 percentstill singleby age16 (PakistanPopulationPlanning Council, n.d., p. 21). Higherfigures for the (singulate) mean age atmarriageare obtainedin the 1974 Retro-spective Survey: 16.5 years for women,24.9 for men. See UK Ministry of Over-seasDevelopment(1977),p. 58, in whichevidencefor recent increasesis also dis-cussed.Seealso note71.

56. The 1974 RetrospectiveSurveyestimated total fertility at 7.1. See alsonote 18.

57. SeeCain ( 1977) for anempiricalstudyof theeconomiccontributionof chil-dren to a family.

58. Theevidenceon food availabilityand nutrition is reviewed by Chen andChaudhury (1975). Alamgir (1974), p.784, calculatesthat in the 1960s a pro-portion fluctuating between 50 percentand 70 percent of the population fellbelow a "minimum" level of living corres-pondingto a daily intakeof 2,100caloriesand 58 gm protein. (On the trend in realwages,seenote 23.)

59. Information from L. C. Chen( personal communication). Cause-of-death data are collected as part of thestatistical program of the Cholera Re-search Laboratory-in Matlab thana, buthave not yet been adequatelyanalyzed.See Curlin, Chen, and Hussain (1976),Table 7 for the Matlab data as presently

W. Bri4n Arthur / GeoffreyMcNicoU

classified. National cause-of-deathesti-matesgiven by Robinson (1967), p. 31,for about 1960 appear to substantiallyoverstate the present contribution ofmalaria and maternal mortality, and tounderstatediarrhealdiseases.

60. A recentCholeraResearchLabo-ratory survey in a rural area on knowl-edge of disease transmissionfound, forexample, that less than 5 percentof re-spondentsbelieved that spread of diar-rheaanddysenterywasrelatedto pollutedwater or' bad food (Cholera ResearchLaboratory,1977, p. 23). In 1969 it wasestimated that two-thirds of unions inrural Bangladeshhad no clinic or dispen-sary; on averagetherewasone doctorper50,000 persons,one paramedicalworkerper5,000(Awan, 1969,pp.73-85).Muchhealthcareis in the handsof homeopaths,hakirns,and other traditional healers.

61. Cain (1978). As could be ex-pected, infant mortality shows compara-tively little class difference, since infantsareaffordedfairly uniform nutritional andimmunological protection by breastfeed-ing.

62. See Jahan (1975) for an over-view of sex differentiaton in the labor-force and occupational discriminationfaced by women. T. Abdullah (1974)gives useful descriptive material onwomen's place in society, their duties,marriage,purdah,and so on. For a widebibliographyon womenin BangladeshseeGermain (1977).

63. See Chen and Ghuznavi (n.d.)and T. Abdullah and Zeidenstein(1977)for interestingcasematerial on the threatof divorce and death of children andhusband.In the 1974 RetrospectiveSur-vey, the proportion of divorced amongever-marriedwomenpeaksat 3.5 percentat ages20 to 24, tailing off in later years.

64. The mortality regimesthatunder-lie the averagedurationof marriageshownin Figure 3 imply that in the presentdecadeabout 17 percentof ever-marriedwomenwould be widowed by age35 and30 percentby age 45; in the 1930s thecorresponding proportions would havebeen33 percentand 50 percent.

73

65. On this practicesee,for example,Bertocci (1970),p. 124.

66. T. Abdullah (1974).

67. Urbanfertility is 7 percentlowerthan rural; the greatest fertility differ-ence among divisions is 5 percent (UKMinistry of OverseasDevelopment,1977,p. 79). Cain and DeVries-Baastiens(1976),tables5 and6, assemble1968-69Impact Survey data indicating almost nofertility differencesbetweenfamilies with"adequate"and "inadequate"living con-ditions,or betweenfamilies with andwith-out agricultural land. For similar findingson landholding differentials see Aird(1956), pp. 271-278, and Stoeckel andChoudhury(1973),p. 123-thelatter giv-ing "desired" rather than attainedfamilysize data. A small positive differential byeducationat its lower levels is found inthe 1974 RetrospectiveSurvey (UK Min-istry of Overseas Development, 1977,chapter5). On urbanmiddle-classfertilitysee Duza (1974), pp. 69, 81, for somerelevantsurveydata.

68. Basedon figures from the Matlabregistration area in Comilla district, theonly sourceof data on year-to-yearvari-ations.The crudebirth rate averaged44.8per thousand over 1966-76, excluding1972 and 1975; birth rates for the lattertwo years were 41.8 and 27.6 (CholeraResearchLaboratory,1977,Table 1).

69. The estimateof meanlactationalamenorrheain Bangladeshis given byChen et aJ. (1974), p. 296. The fertilityimpact is calculated using the formulasand ceteris paribus assumptionsof Bon-gaarts (1978). If lactation dropped toUS levels, fertility could be raised 75percent.

70. SeeKnodel (1977).

71. A more significant effect of poornutrition is to delay menarche. Recentresearchsuggests,however, that betternutrition in Bangladeshwould at mostraise fertility by around 5 percent. SeeBongaartsand Delgado(1977) and Mos-ley (1977). It has been speculatedthatthe apparent recent increase in age atmarriagereflectsthe delayedmenarcheof


girls who experiencedthe food crises ofthe early 1970s,since onsetof menarcheoften triggersmarriage(CholeraResearchLaboratory,1977,p. 40).

72. Potentill1 fertility lost by deathof thewife by definition doesnot enterthemeasureof total fertility, henceis not con-sideredhere.

73. Chenet aI. (1974),p. 296.

74. SeeAhal ( 1967),p. 61, and UKMinistry of OverseasDevelopment(1974).An earlier village demographicstudy byAird (1956), p. 157, concludedthat be-tween 96 and 98 percent of women areinitially capableof procreation.

75. Secondarysterility shows someuniformity across different populations.Henry's data for European populations(perhapsa lower bound for Bangladesh)average 3 percent of married womensterile at age 20, 6 percentat age25, 10percentat age 30, 16 percentat age 35,and 31 percentat age 40 (Henry, 1961,p.85).

76. Sirageldin, Hossain, and Cain(1975),p. 6. A slightly higherpercentage(6.4) had ever used contraception.In-duced abortion is thought to be almostnonexistent(Chenet al., 1974, p. 279).

77. Sirageldin, Hossain, and Cain(1975), p. 24, concludefrom analysisofthe Impact Survey: "Even if we take allcurrentuseandexpectedfuture use,basedon reportedintentions,this total expecteddemandwas and is not large enough togenerate the desired fertility reductionsin Bangladesh."

78. The experiment,begun in 1975,covers a population of 125,000 with asimilar-sized control group. A user rateof 10 percent might lead to a drop ofperhaps5 points in the birth rate, say toaround40 per thousand.It would be diffi-cult, however,to replicatesucha programon a wide scale. See A. R. Khan andHuber (1977) and Cholera ResearchLaboratory (1977), pp. 36-J7.

79. Duza (1974), p. 82, cites datafrom a 1974 survey of male villagers inrural Chittagong: 86 percentof respond-

ents thought the country was now inworsecondition than in the past,and 80percentconsideredtheir own condition tobe less happy than that of their fore-fathers.

80. The argument is developedbyCain (1978), who finds evidenceof the;'crisismanagementfunction" providedbydepthin the householdlabor force.

81. The termis Abu Abdullah's(per-sonalcommunication).

82. Of ever-marriedwomen in their40sin 1969,13 percenthad no living sons(calculated from Impact Survey data).The correspondingpercentagefor womenat this lowest economic level may wellexceedtwice that figure.

83. From this description it mightseem that no increasein female maritalage could be expected.But forces thatwould tend to raise this age can be de-tected: lesseningstrengthof the patriar-chal family as assetsare diluted (controlof a son's behavior is maintained bypromise of inheritancerather than filialpiety) and gradual raising of women'sstatusIn the family.

84. Baldwin (1977) contrastsa pop-ulation projectionassumingconstantmor-tality and fertility at current levels withone in which there are an additional500,000 disaster fatalities annually, andshowsthat the populationattainedby theyear2000In theformercase( 156million)is reachedonly six yearslater in the lattercase.

85. India's industrial productionaveraged6.1 percentannualgrowth over1951-73 (Veit, 1976,p. 75).

86. The index of industrial produc-tion in Bangladeshon a base1970= 100hasrecordedthefollowing recentchanges:1974, 90; 1975, 81; 1976, 87; 1977, ff1(Statistical Pocket Book uf Banglodesh1978,p. 178).

87. In a recentsurveyof union COUD-

cil members,M. Rashiduzzaman(personalcommunication) found evidence of a"new" elite beginning to emergein addi-tion to the landed gentry-professional

W. Brian Arthur / Geoffrey McNicoU

people,local contractors,and other smallbusinessmen.

88. Basedon Matlab data given byChaudhuryand Curlin (1975) and Mir-anda (1977), on averagein a village of500 people, perhaps 5 persons a yearwould move to a town and 2 or 3 return.The ＱＹＶＸｾＹ National Impact Surveyfound that 25 percentof married womenvillagers had ever visited a town, and 15percenthaddoneso in the past12months(Pakistan Population Planning Council,n.d., p. 36). The fli\1res for men would,of course,be greater.

89. SeeBangladeshRural Advance-ment Committee (1976) and Chen andGhuznavi(n.d.).

90. Using Cain's data (1977), thiswould happen if rural wagesfell by 30percent.

91. For an account of family plan-ning program activities see Bangladesh,PopulationandNutrition ProjectsDepart-ment(1975).A broadcriticism of thepro-gramis containedin Demeny (1975).

BIBLIOGRAPHY

75

92. See, for example, M. Ahmed(1977a,1977b)and Haqueet al. (1977).The latter gives a brief report on the so-called Rangpur Self-reliant Movement,which achieved remarkable and widelypublicized results in organizing peasantsfor a variety of developmentobjectives(including population stabilization) in asingle village (Kunjipulcur) in Rangpurdistrict, and which has subsequentlyspreadto othervillagesin the area.

93. The Swanirvar movement takesmuch of its impetusfrom the isolatedex-amples mentioned in the previous note.SeeAlamgir (1977), chapter5, for a de-scription and skepticalassessment.

94. SeeBangladesh,PopulationCon-trol and Family PlanningDivision (1976),p.12.

95. Signs of interest in this generalpolicy approachinclude the discussionof"compulsory cooperatives" in the lastmonthsof the Mujib administration,andthe still-deTelopingSWaWrvarmovement,mentionedearlier.

Abdullah,Abu. 1976."Land refonnandagrarianchangein Bangladesh."Bangla-deshDevelopmentStudies4, no. 1; 67-99.

---, MosharaffHossain,and RichardNations.1974. SIDA/fLO ReportonIntegratedRural DevelopmentProgramme,Banglcuksh.Dacca.

-. 1976."Agrarianstructureandthe IRDP-preliminaryconsiderations."BanglcukshDevelopmentStudies4, no. 2: 209-266.

Abdullah, TahrunessaAhmed. 1974. "Village women as I saw them." Mimeo,The Ford Foundation,Dacca. (Originally publishedin 1966.)

---, and Sondra Zeidenstein.1977. "Rural women in development."InRole of Women in Socio-EconcmicDevelopment in Bangladesh.Dacca:BangladeshEconomicAssociation.

Adnan, Shapan.1976. "Land, power and violence in Barisal villages," DaccaUniversityVillage StudyGroup,Working Paperno. 6.et al. 1975."The preliminary findings of a social andeconomicstudyof four Bangladeshvillages." Dacca University Studies23 (June):111-127.

---. 1977. "Differentiation and classstructurein village Shamraj,"DaccaUniversity Village StudyGroup,Working Paperno. 8.


Afzal, Mohammed.1967. "The fertility of East Pakistanmarried women." InStudiesin the Demographyof Pakistan,ed. W. C. Robinson.Karachi:PakistanInstituteof DevelopmentEconomics.

Ahmad, NaBs. 1956. "The patternof rural settlementin EastPakistan."Geo-graphicalReview46,no. 3: 388-398.

--'-. 1968. An EconomicGeographyof East Pakistan, 2nd ed. London:Oxford UniversityPress..

Ahmed, A. S. M. Mohiuddin. 1962. "The population of Pakistan: Past andpresent:'Ph.D.dissertation,DukeUniversity.

Ahmed, Manzoor. 1977a."BRAC: Building humaninfrastructureto serve therural poor." Mimeo, InternationalCouncil for EducationalDevelop-ment,Essex,Conn.

---. 1977b."The SavarProject: Meeting the rural healthcrisis in Bangla-desh:' Mimeo, InternationalCouncil for EducationalDevelopment,Essex,Conn.

Ahmed,Noazesh.1976.DevelopmentAgricultureof Bangltulesh.Dacca:Bangla-deshBooksInternational.

Aird, John S. 1956. "Fertility levels and differentials in two peasantvillages."Ph.D.dissertation,Universityof Michigan.

Alamgir, Mohiuddin. 1973. "Approaches towards researchmethodology onproblems of urbanization in Bangladesh."ResearchReport no. 15(newseries),BangladeshInstituteof DevelopmentEconomics,Dacca.

---. 1974. "Someanalysisof distribution of income,consumption,savings,andpovertyin Bangladesh:'BangltuleshDevelopmentStudies2: 737-818.

---. 1975. "Someaspectsof Bangladeshagriculture: Review of perform-ance and evaluationof policies." Bangltulesh DevelopmentStudies3:261-300.1977. Bangladesh:Acmeof Below P(1)erly Level Equilibrium Trap.Dacca:BangladeshInstituteof DevelopmentStudies.

Awan, Akhtar. 1969. The Systemof Local Health Servicesin Rural Pakistan.Lahore:PublicHealthAssociationof Pakistan.

Baldwin, C. Stephen.1977. "Catastrophein Bangladesh:An examinationofalternative populationgrowth possibilities1975-2000:'Asian Survey17: 345-357.

Bangladesh.Bureau of Statistics. 1977. Yearbookof Agricultural StatisticsofBangladesh1975-76.Dacca.

---. Bureauof Statistics.1978.StatisticalPocketBookof Bangladesh1978.Dacca.

---. Population and Nutrition Projects Department 1975. "The popu-lation programof the Governmentof Bangladesh:A sectorreview."Dacca.PopulationControl and Family PlanningDivision. 1976. "An outlineof the NationalPopulationPolicy of Bangladesh."Mimeo.

BangladeshRural AdvancementCommittee(BRAC). 1976. "Pachbarol'sdesti-tute women'scooperative."BRAG NewsletterI, no. 6 (November-December).

Bertocci,PeterJ. 1970."Elusivevillages: Socialstructureand communityorga-

W. Brian Arthur / Geoffrey McNicoU 77

nization in rural East Pakistan."ph.D. dissertation,Michigan StateUniversity.1976. "Rural developmentof Bangladesh."In Rural DevelopmentinBangladeshand Pakistan,ed. Robert D. Stevens,Hamza Alavi, andPeterJ. Bertocci.Honolulu: UniversityPressof Hawaii.

Bhatia,B. M. 1967.Faminesin India: A Studyof SomeAspectsof the EconomicHistory of India (1 86Q-1965),2nded.Bombay:Asia PublishingHouse.

Blacker, J. G..c. 1977. "Dual record demographicsurveys: Are-assessment."PopulationStudies31: 585-597.

Bongaarts,John. 1978."A frameworkfor analyzingthe proximatedetenninantsof fertility." This volume.

---, and HernanDelgado.1977."Effects of nutritional statuson fertility."In Proceedingsof the International Population ConferenceMexico,vol. 1. Liege: InternationalUnion for the Scientific Study of Popula-tion.

Cain, Mead.1977."The economicactivitiesof childrenin a Bangladeshvillage."Populationand DevelopmentReview3, no. 3 (September):201-227.

---. 1978. "Economic class, economic mobility and the developmentalcycle of households:A casestudy in rural Bangladesh."Paperpre-sentedto theannualmeetingof thePopulationAssociationof America,Atlanta,Georgia.

---, and Woutje DeVries-Baastiens.1976. "Householdstructuresand fer-tility in Bangladesh."Paperpresentedat the annualmeetingof thePopulationAssociationof America,Montreal,28 April-l May.

Chaudhury,Rafiqul Huda,andGeorgeC. Curlin. 1975."Dynamicsof migrationin a rural areaof Bangladesh."BangladeshDevelopmentStudies3,no. 2: 181-230.

Chen,L. C., ed. 1973. Disaster in Bangladesh:Health Crisis in a DevelopingNation. New York: Oxford University Press.et al. 1974. "A prospectivestudy of birth interval dynamicsin ruralBangladesh."PopulationStudies28: 277-298.

----.--, andRafiqul Huda Chaudhury.1975."Demographicchangeand foodproduction in Bangladesh,1960-74." Population and DevelopmentReviewI, no. 2 (December):201-227.

Chen,Marty, andRuby Ghuznavi.n.d. "Women in food-for-work: The Bangla-deshexperience."Unpublishedpaper.

CholeraResearchLaboratory.1977.AnnualReport1977. Dacca.Curlin, G. T., L. C. Chen,and S. B. Hussain.1976. "Demographiccrisis: The

impactof the Bangladeshcivil war (1971) on births and deathsin arural areaof Bangladesh."PopulationStudies30: 87-105.

DaccaUniversity Village StudyGroup. 1975."Social structureand implicationsfor resourceallocationin a Chittagongvillage." Working Paperno. 3.

Davis, Kingsley. 1951.The Populationof India and Pakistan.Princeton:Prince-ton UniversityPress.

Demeny,Paul. 1975. "Observationson population policy and population pro-gram in Bangladesh."Population and DevelopmentReview I, no. 2(December): 307-321.


Duza,M. Badrod.1974."Cultural consequencesof populationchangein Bangla-desh."Paperpreparedfor the seminaron Cultural ConsequencesofPopulationChange,14-17August 1974, Bucharest,Romania.

Eaton,JosephW., and Albert J. Mayer. 1953."The socialbiology of very highfertility amongthe Hutterites."Human Biology 25: 206-264.

Elahi, K. M. 1974."Spatialdistributionof populationstructuresin Bangladesh."In Studiesin BangladeshGeography,ed.A. F. M. Kamaluddin.Dacca:BangladeshNationalGeographicalAssociation.

er-Rashid,Haroun. 1967. EastPakistan:A SystematicRegionalGeographyandIts DevelopmentPlanningAspects,2nded. Lahore: GhulamAli.

Faaland,Just,and J. R. Parkinson.1976.Bangladesh:The TestCasefor Devel-opment.London:Hurst.

Farouk,A. 1976.TheVagrantsof DaccaCity: A SocioeconomicSurvey.Dacca:Bureauof EconomicResearch,DaccaUniversity.

Gennain,Adrienne. 1977. "Notes on indicators of women'sstatusand roles,"Paperpreparedfor the Workshopon Country Reportson Populationand Development,Bergen,Norway, September1977.

Haq, Mahbubul. 1963. The Strategyof EconomicPlanning: A CaseStudyofPakistan.Karachi: Oxford University Press.

Haque,Wahidulet a!. 1977."Towardsa theoryof rural development."Develop-mentDialogueno. 2.

Henry, L. 1961."Somedataon natural fertility." EugenicsQuarterly 8: 81-91.Huq, M. Ameerul,ed. 1976.Exploitationandthe Rural Poor: A Working Paper

on the Rural Power Structure in Bangladesh.Comilla: BangladeshAcademyfor RuralDevelopment.

Islam, A. K. M. Aminul. 1968."Changesin croppingpatternsin EastPakistan."OrientalGeographer12: 1-25.

---. 1974.A BangladeshVillage, Conflict and Cohesion:An Anthropologi-cal Studyof Politics. Cambridge,Mass.: Schenkman.

Islam, Nurul. 1974."The stateandprospectsof the economy."In The EconomicDevelopmentof Bangladeshwithin a SocialistFramework,ed.E. A. G.RobinsonandK. B. Griffin. London: Macmillan.

Jaban,Rounaq.1975."Women in Bangladesh,"In Womenfor Women:Bangla-desh1975. Dacca:UniversityPress.

Khan, Akhter Hameed.1971.Tour of TwentyTOOnas,21-12-1970to 28-1-1971.Comilla: BangladeshAcademyof RuralDevelopment.

Khan, Atiqur R., and DouglasH. Huber. 1977."Householddistributionof con-traceptivesin Bangladesh-ruralexperience."Paperpresentedat theRegional Conferenceon Village and HouseholdAvailability of Con-traceptives,Tunis,27-30March1977.

Khan, Masihur Rahman,1972a."Bangladeshpopulationduring the first FiveYear Planperiod (1972-1977):A guestimate,"ResearchReportno. 6(newseries),BangladeshInstituteof DevelopmentEconomics,Dacca.

---. 1972b."Migration within andacrossthe boundariesof Eastand WestPakistan,1901-61," Ph.D. dissertation,Australian National Univer-sity, Canberra.

---, andAminur RohmanKhan. 1975."Demographicprocessesin Bangla-desh." Country background paper for the Seminar on Population

W. Brian.Arthur / Geoffrey McNicoli 79

Policy, Bangladesh,15-21 May 1975. Population Study Centre,BangladeshInstituteof DevelopmentStudies,Dacca.

Knodel,John.1977."Breast-feedingandpopulationgrowth." Science198: 1111-1115.

Majlls, Daud. 1977. "A basketful of hope." Far EastemEconomic Review.(23September):ｾＲＵＮ

Majumdar,R. C., ed. 1943-48.History of Bengal,2 vols. Vol. 1, Dacca,1943;Vol. 2, Calcutta,1948.

McCord, Colin. 1976."What's the useof a demonstrationproject?"Mimeo, TheFordFoundation,Dacca. (Reportno. 45)

Miranda, Armindo. 1977. "Les migrations au Bangladesh."Population no. 2:448-459.

Morris, Morris D., and Burton Stein. 1961. "The economichistory of India: Abibliographicessay."Joumalof Economic History21: 179--207.

Mosley, Henry. 1977. "The effectsof nutrition on natural fertility." Paperpre-sentedat the Seminaron Natural Fertility, Institut national d'Etudesdemographiques,Paris,March1977.

PakistanPopulationPlanning Council. n.d. National Impact Survey Report.Lahore,c. 1970.

Papanek,Gustav F. 1967. Pakistan'sDevelopment:Social Goals and PrivateEnterprise.Cambridge,Mass.: HarvardUniversity Press.

Potter,J. E. 1977."Problemsin using birth-historyanalysisto estimatetrendsinfertility." PopulationStudies31: 335-364.

Rahman,A. T. R. 1962.BasicDemocraciesat theGrassRoots.Comilla: PakistanAcademyfor Village Development.

Ranis,Gustav.1974."Brief reHectionson the central issuesof policy in Bangla-desh." BangladeshDevelopmentStudies2: 839--856.

Raper,Arthur F. 1970. Rural Developmentin Action: The ComprehensiveEx-perimentat Comilla, East Pakistan. Ithaca, N.Y.: Cornell UniversityPress.

Robinson,WarrenC. 1967."Recentmortality trendsin Pakistan." In Studiesinthe Demographyof Pakistan,ed. W. C. Robinson.Karachi: PakistanInstituteof DevelopmentEconomics.

Sirageldin,Ismael,MonowarHossain,and Mead Cain. 1975. "Family planningin Bangladesh:An empirical investigation." BangladeshDevelopmentStudies3, no. 1: 1-26.

, D. Norris, and M. Ahmad. 1975. "Fertility in Bangladesh:Factsandfancies." PopulationStudies29: 207-215.

Sobhan,Rehman.1968. Basic Democracies,Works Programmeand Rural De-velopmentin EastPakistan.Dacca:Oxford UniversityPress.

Stoeckel,John,andMoqbul A. Choudhury.1973.Fertility, Infant Mortality andFamily Planningin Rural Bangladesh.Dacca:Oxford UniversityPress.

Tepper,Elliot. 1970."Rural developmentand administrationin EastPakistan."Ph.D.dissertation,DukeUniversity.1976. "The administrationof rural refonn: Structuralconstraints."InRuralDevelopmentin BangladeshandPakistan,ed.RobertD. Stevens,Hamza Alavi, and Peter J. Bertocci. Honolulu: University PressofHawaii.


Thomas,John.1968. "The Rural Public Works Programmeand EastPakistan'sdevelopment."Ph.D. dissertation,HarvardUniversity.

UK Ministry of OverseasDevelopment.1977. Reporton the 1974 BangladeshRetrospectiveSuroeyof Fertility and Mortality, ed. J. G. C. Blacker.LondonandDacca.

United Nations.GeneralAssembly.1976."Requestfor the inclusionof a supple-mentary item in the agendaof the Thirty.first Session:Questionofunilateral diversion of waters of the international river Ganges,incontraventionof all internationallaws andregulationsand tradi'.ionalusagesand in violation of solemnpledgeson the useof suchwaters."Letter dated26 August 1976 from the PennanentRepresentativeofBangladeshto the United Nations,addressedto the Secretary-General.UNdocumentA/31;195/Add.1.NewYork.

Veit, LawrenceA. 1976. India's SecondRevolution:The Dimensionsof Devel-opment.NewYork: McGraw-Hill.

Wood, GeoffreyD. 1976."ClassdiHerentiationandpowerin Bondokgram:Theminifundist case." In Exploitation and the Rural Poor: A WorkingPaper on the Rural Power Structure in Bangladesh,ed. M. AmeerulHuq. Comilla: BangladeshAcademyfor Rural Development.

an analytical survey of population and …an analytical survey ofpopulation and development in...

Documents