estimating the importance of plantations

7/29/2019 Estimating the Importance of plantations

1/6

Economic History Association

Estimating the Importance of the Plantation System to Southern Agriculture in 1880Author(s): Nancy VirtsSource: The Journal of Economic History, Vol. 47, No. 4 (Dec., 1987), pp. 984-988Published by: Cambridge University Press on behalf of the Economic History AssociationStable URL: http://www.jstor.org/stable/2122042

Accessed: 19/08/2010 18:19

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=cup.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

Cambridge University Press andEconomic History Association are collaborating with JSTOR to digitize,

preserve and extend access to The Journal of Economic History.

http://www.jstor.org
http://www.jstor.org/stable/2122042?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=cuphttp://www.jstor.org/action/showPublisher?publisherCode=cuphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2122042?origin=JSTOR-pdf


2/6

Estimating the Importance of the PlantationSystem to Southern Agriculture in 1880NANCY VIRTS

Althoughthe importanceof the plantation ystem to antebellum outhernagricultureis widely acknowledged, there is less consensus on the role of the plantationsystemafter the Civil War. Roger Ransom and Richard Sutch conclude in their work onpostbellumsouthernagriculture,One Kindof Freedom, that even thoughlandowner-ship remainedconcentratedand largelandownerscontinued to exercise economic andpolitical power, by 1880 the plantationsystem had "ceased to exist."'"They estimatethat largefarms(farmswithmore than 200 acres) dependenton wage labormade up lessthan 1percentof all farms in the CottonSouth.2Using theirdata, theirmethods, andasampleof the 1880 census manuscriptsof populationand agriculture,I estimate thatthese farmsproducedabout 14percentof the cotton growninthe Cotton South. In 1860plantationswith greaterthan 50 slaves were about4 percent of the farmsin the Southand produced about 32 percent of the South's cotton.3 The apparent decline inimportanceof large cotton-producingarms after emancipation upports he theory thateconomies of scale on slave plantationswere associated with slavery rather thanmethods of large-scaleproductionand could not be duplicatedwith free labor.Whilemy results indicatethat the importanceof large armsdependenton wage laborwas limited n the postbellumSouth,it is misleading o conclude on this basis alone thatthe plantation system had ceased to exist by 1880. Many large landholdings n thecotton-growingareas afterthe Civil Warwere workedby tenant not wage labor. Theselandholdings,commonlyknown as tenantplantations,were managedas a single uniteven thoughactual cultivationwas done by tenants, most of whomwere sharecroppers.The similaritiesbetween managementand productionmethods used on tenantplanta-tions and antebellum slave plantations suggest that tenant plantations should beincludedin measures of the plantationsystem. Landownerswere actively involved incultivating heirland, supervising he planting,andharvestingandmarketing he crop.4

Journalof EconomicHistory, Vol. XLVII, No. 4 (Dec. 1987).C EconomicHistoryAssociation.All rights reserved. ISSN 0022-0507.The author s lecturerof Economics, CaliforniaState University,Northridge,CA 91330.This note developed from my Ph.D. dissertationat UCLA. I would like to thank the chairmanof my dissertationcommittee, KennethSokoloff, the membersof the committee, BenjaminKleinand MaryYaeger, as well as ArmenAlchian, HaroldDemsetz, Stanley Engerman,Robert Fogel,members of the Sloan Workshopat UCLA and participantsat the All UC Economic HistoryConferenceSpring 1984for their commentsand support.A special thanksalso to Lee Alston,Roger Ransom and RichardSutch for sharingdata with me as well as many helpfulcomments.Financial upport rom heAll UC EconomicHistoryGroup, he EarhartFoundation nd the SloanFoundation s acknowledged.

1Roger Ransom and Richard Sutch, One Kind of Freedom: The Economic Consequences ofEmancipation Cambridge,Mass., 1977),p. 56.2 Ransom and Sutch define a plantationas a farm reporting200acres or morein crops, greaterthan 98 weeks of hired labor and relying on hired labor for at least 60 percent of the laborrequirements.See Ransomand Sutch, ibid., p. 68, appendixG.3The five cotton states are Mississippi, Louisiana, Georgia, South Carolina and Alabama.Ransom and Sutch, ibid., p. 74.4 Fordescriptionsof the administrationf tenantplantations ee RobertBrooks,"TheAgrarianRevolution in Georgia," University of Wisconsin, Bulletin,No. 639 (Madison, 1914) pp. 1-77;C. 0. Brannen, "Relation of Land Tenureto PlantationOrganization,"U. S. DepartmentofAgriculture, Bulletin, No. 1269 (Oct. 1924); and T. J. Woofter, Landlord and Tenant on the CottonPlantation, (WashingtonD.C., 1936).

984


3/6

Postbellum Plantation System 985Tenantson these plantations ometimesworked ointly under he directionof the owneror his agent. It was not uncommon or plantations,even as late as 1880, to be workedby squads of laborers arger han family groups.5On some plantations he entire tenantlabor force sometimes worked jointly for certain activities in a system known as"through-and-through.6No statistics on tenantplantationswere collected by the Officeof the Census in 1880.Census enumeratorsn 1880were instructed o reporteach tenantfarm as an indepen-dent unit even if it was part of a largerunit. Therefore,Ransomand Sutch'sestimate ofthe numberof largefarms dependenton wage labor s acceptableonly as a lower-boundestimate of the importanceof the plantationsystem.I will establish that the method Ransom and Sutch used to "clean" the sample offarmsbiases their estimate of the importanceof the plantation ystem downward.Theirprocedureremoved a significantpercentage of the large farms in the sample. Beforeestimating he distribution f farms by size, Ransomand Sutch deleted from the samplethose farms with missing or ambiguous nformation n variablessuch as the race of thefarmoperatorand tenureof the farm, and those farms where the laborinputappearedinadequate. The reported amount of labor was judged inadequate if no labor wasreportedor if the ratio of land in crops to laborwas greater han 200 acres.7The impactof Ransom and Sutch's restrictionson the dataset as a whole and for largefarms(200acres andabove) is shown in Table 1.Although he effect on the data set as a whole is not large, the effect on large farms sconsiderable.Theirproceduredeletes only 11percent of the total numberof observa-tions in the sample, and the structureof the data set with respect to race andtenureisunchanged. n contrast, over 40 percentof the large arms n the total sample ail to meetthe imposed restrictions.In the Cotton South, 47 percent of the farms with greater han200acres were deletedby Ransom and Sutch's procedure.Overhalfof the farmswereomittedbecause of inadequate abor input. Whilethis procedurecan be justifiedwhenthe major interest is estimatinglabor productivity,it will bias any estimate of theimportanceof a particular ize of farm.Giventhe numberof largefarmsaffected,the source of these problems s a matterofinterest. It is possible that a disproportionate umberof large farms were deleted fromthe samplebecause of problemswith the reportingof hiredlabor.Census enumeratorsin 1880experiencedso much difficulty ollecting nformation n hired aborthatthedatawere not compiled in the aggregate census report.8Enumeratorswere instructedtocollect informationon the total wage bill, the man-weeksof white laborhired,and theman-weeksof "colored" labor hired,butthese instructionswere frequentlynot carriedout. At times enumerators ailed to recordany information n hired labor.Moreoftenthe wage bill appearsto have been correctlyrecorded while the weeks of laborhiredwere not.9

5For a more detailed descriptionof the squad system see Ralph Shlomowitz, TransitionFromSlave to Freedman: Labor Arrangements in Southern Agriculture, 1865-1870 (Ph.D. dissertation,University of Chicago, 1979); and Ralph Shlomowitz, "Origins of SouthernSharecropping,"Agricultural History, 53 (July 1979), pp. 557-75.

6 See Brooks "AgrarianRevolution,"p. 66. Merle Prunty, "The Renaissanceof the SouthernPlantation,"GeographicalReview, 45 (Oct. 1955), p. 468. Other references to the through-and-through system in Mississippi are Frank Welch, "The Plantation Land Tenure System inMississippi,"MississippiAgricultural xperiment tation,Bulletin,No. 385(June 1943),p. 49; andHarold Pedersonand ArthurRaper,"The Cotton Plantation n Transition,"MississippiAgricul-tural ExperimentStation, Bulletin, No. 508 (Jan. 1954),p. 16. Welch notes that some Mississippiplantationshad been worked on the through-and-throughystem for long periods of time.7This procedure s described n Ransom and Sutch, One Kind of Freedom, appendixG.8Ibid., fn. 18, p.370.9Enumerators often recorded he numberof weeks duringwhich laborwas hired nsteadof the


4/6

986 VirtsTABLE 1FARMS WITH MISSING INFORMATION FOR SELECTED VARIABLES IN THERANSOM AND SUTCH SAMPLE

Farms with Greater thanTotal Sample 200 Acres in CropsNumber Percent Number Percent

No Labor 535 5% 0 0%More than 200 acres per 157 1 47 25workerMissing information onOperatora 363 3 26 14Tenure 80 1 1 1Laborb 360 3 2 1Total deleted 1,211 11 76 41Total farms 11,202 100 185 100

a Includes cases where racial information was missing and cases where farm owners could not belocated in the population census.b Includes cases where operator was listed as operating more than one farm and cases whereoperator resided on a farm other than the one he operated.c The total deleted does not equal the sum of each column because some observations havemultiple errors.Source: Computed from the Roger Ransom and Richard Sutch sample, One Kind of Freedom: TheEconomic Consequences of Emancipation (Cambridge, Mass., 1977), appendix G.

It is possible that farms with high land-to-labor ratios may have used hired laborwhich was not counted because of the carelessness of enumerators or their misunder-standing the instructions. But there is evidence to suggest that the absence of labor onlarge farms was due to other factors. As shown in Table 2, the percentage of farmsgreater than 200 acres with a land-to-labor ratio greater than 200 acres was almost twiceas high in the Cotton South than in the rest of the South.Because almost the same percentage of farms in both areas were in districts where thewage data were classified as reliable by Ransom and Sutch, it seems that the quality ofenumerators did not vary greatly among areas.' 0 None of the rice and sugar plantationsin the sample was deleted as a result of insufficient labor even though these crops weremore labor intensive than cotton." Thus it seems likely that the underreporting of labortotal numberof weeks of hired abor. Afterexamining he dataRansomandSutch concludedthatenumeratorswere in general consistent in the way they recorded wage data. They classifiedenumeration istrictsas eitherreliableorunreliablebased on the averageweeklywage obtainedbydividingwage bill by man-weeks of laborhired. Districts where the averageweekly wage wasgreater han $4were classifiedas unreliable.However, even in cases where only the wagebill wasreliablyrecorded t is possible to estimate the amountof laborhired. To estimate the amountoflabor hired in these cases, average weekly wage rates by county were calculated using thosedistrictscoded as reliable. Forfarms n districtscoded reliableand with information n weeks ofhired abor,labor was recordedas the sum of man weeks of whiteand blacklabordivided by 52.For farmswithonlythewagebill recorded aborwas estimatedas the total wagebilldividedby theaverageweekly wagefor thatcounty dividedby 52. For moredetaileddescription ee RansomandSutch, One Kind of Freedom, appendix G.

10Forfarmsgreater han200 acres in the Cotton South 58 percentwere in enumeration istrictscodedreliable; n the rest of the South 53 percentof the farmswere in enumeration istrictscodedreliable.l See LewisC. Grayet al., "FarmOwnership nd Tenancy,"U. S. Department f Agriculture,Yearbook 1923 (Washington D.C., 1924), pp. 530-32.


5/6

Postbellum Plantation System 987TABLE 2CLASSIFICATION OF FARMS WITH OVER 200 ACRES IN THE RANSOM ANDSUTCH SAMPLE

Cotton South Rest of SouthNumber Percent Number Percent

Plantations 46 38% 39 61%Excluded from plantation category 19 16 5 8because of labor compositionExcluded from sampleMore than 200 acres per worker 37 31 10 16Missing information on operatorb 17 14 9 14Ambiguous information on 0 0 9 14tenure'Other 2 2 0 0

a Farms were excluded from the plantation category because they hired less than 98 weeks of laboror less than 60 percent of their work force was hired.b Farms where racial information was illegible and where farm owners could not be located in thepopulation census.c Farms where the operator was listed as operating more than one farm or where the operatorresided on a farm other than the one he was listed as operating.Source: Computed from the Roger Ransom and Richard Sutch sample, One Kind of Freedom: TheEconomic Consequences of Emancipation (Cambridge, Mass., 1977), appendix G.

was somehow related to the production of cotton. While some enumerators probably didmiscount labor, the systematic nature of their errors supports the notion that theprimary factor was the confusing interdependence of production among tenants.These landholdings were often managed as a single unit even though the labor wassupplied by tenants. On some plantations tenants worked together under the landlord'sdirection to perform certain tasks. Because landlords were actively directing cultivationon many tenant plantations, it is possible that some enumerators were reluctant to treattenant plots as individually operated farms.

One enumerator noted on the census manuscripts that he recorded the total amountof capital, livestock, acreage, and land value for tenant plantations in his entry for theowner-cultivator's farm. Entries for tenants working the plantation contained only thename of the tenant farmer, his tenure status, and crop acreage.'2 Large owner-cultivatedfarms in this district appear to have insufficient labor because they were worked withtenant, not wage labor. Other statements by census officials and the geographicalpattern of large farms on which labor was underreported in the Ransom and Sutchsample suggest that other enumerators followed similar procedures when reportingtenant plantations.These results suggest that one can get a better idea of the quantitative importance ofthe plantation system by estimating cotton production by the size of farm withoutreference to type or amount of labor used. If only those farms with no productioninformation are deleted, all farms over 200 acres remain in the sample. Some of thesefarms may have been tenant plantations. These results are presented in Table 3.Farms with 200 or more acres account for 24 percent of cotton produced in 1880,about 75 percent of the share of cotton produced by large slave plantations in theantebellum period. This figure should be considered a lower-bound estimate of thecotton produced by the plantation system because some enumerators did follow

12 For example see the records of enumerators romConcordiaParishLouisiana.


6/6

988 VirtsTABLE 3COTTON PRODUCTION BY FARM SIZE, 1880

TenureFarm Size Own Rent Share50 acres or less 19% 6% 15%51 to 100 acres 15 3 4101tol99acres 11 1 1200 acres or more 23 1 0Total 68 11 20Source: Computed from the Roger Ransom and Richard Sutch sample, One Kind of Freedom: TheEconomic Consequences of Emancipation (Cambridge, Mass., 1977), appendix G.

instructions and reported the output of tenant plantations on tenant farms. Part of theoutput of tenant plantations was reported with the output of renters and sharecroppers.By adding the percentage of cotton produced on large farms to that produced on smallsharecropped farms it is possible to get an upper-bound estimate of the importance ofthe plantation system. The upward bias caused by including the production ofsharecroppers who were not part of plantations is partially compensated for by notincluding the production of renters who were part of plantations. Although the majorityof plantation tenants were sharecroppers, renters were substantial minorities on someplantations.The actual percentage of cotton produced by the plantation system in 1880 waspresumably between the 24 percent produced on farms with 200 or more acres and 39percent, the percentage of cotton produced on these large farms and sharecropped farmswith 50 acres or less. Both figures are substantially higher than the 14 percent estimatedby Ransom and Sutch to have been produced on cotton farms operated with wage labor.The amount of cotton produced on tenant plantations in the postbellum period waslower than the percentage of cotton produced on slave plantations of similar size in theantebellum period. But the plantation system and the large-scale cultivation of cottonwere still important factors in 1880.

estimating the importance of plantations

Documents