Social Status and Wok Incentives in Shaecoing. The Case of Rual Bangladesh A Note Jon Reiesen * Vestfold College Deatment of Economics and Business Administation P.O. Box 2243, N-3103 Tønsbeg NORWAY Abstact: In ual Bangladesh as elsewhee social concens and influences shae economic decisions. This note illustates how the quest fo social status among smallholdes may exlain why landlods efe landed athe than landless famwokes as thei tenants. The geate cost of losing tenancy fo a landed tenant, stemming fom his socio-sychological avesion to offeing his labo in the casual labo maket, induces him to cultivate the ented land with geate cae and intensity comaed to a landless tenant with no social status to defend. Theefoe, the landlod gets a highe etun if he ents out his land to a landed athe than a landless fame. * The ae is at of the oject Institutions, Social Noms, and Economic Develoment. I have eceived helful comments by Kalle Moene. Financial suot fom the Nowegian Reseach Council is gatefully acknowledged.
1. Intoduction In ecent yeas, economists have etuned to the ole of institutions in economic develoment. By exloing oblems of imefect infomation and missing makets, ogess has been made in exlaining the undelying ationale and consequences of diffeent economic institutions often obseved in agaian economies in the Thid Wold. 1 Recent theoetical liteatue has, howeve, left little oom fo influences aising fom cultual values and social noms. A fulle undestanding of agaian oganization in the Thid Wold, including its vaiation acoss majo egions, thus equies that the social and cultual contexts in which eole oeate must be taken moe seiously. It will be agued in this ae that in ual Bangladesh as elsewhee economic action is constained and shaed by social concens and influences. In aticula it will be shown how the quest fo social status may exlain why landlods have such a stong efeence fo landed athe than landless tenants. About 80 million of the 90 million eole who live in ual Bangladesh deend on agicultual activities fo thei income and emloyment. Thee is, howeve, an exteme scacity of both land and ootunities fo emloyment, and it is fo these two souces of income that eole comete. It is estimated that about 48 ecent of ual households in Bangladesh ae landless, o have so little land that they ae deendent on leasing in land o woking as wage laboes in ode to maintain thei mateial standad of living. and leasing is an imotant featue of agicultue in Bangladesh. About 23 ecent of the total cultivated land is famed unde vaious tenuial aangements. The most common tenancy aangement is shaecoing, which accounts fo nealy 90 ecent of the total 1 The main message fom this liteatue is that diffeent institutional aangements and contacts emege as substitutes fo missing cedit and insuance makets in an envionment of evasive isks, infomation asymmety, and moal hazad. Fo examles and oveviews of the liteatue, see Stiglitz 1988 and Badhan 1989. 1
leased land Januzzi and Peach, 1980. The cometition fo land to shaeco in Bangladesh is fiece. Emiical evidence eveals that it is almost imossible fo a landless fame to obtain land on a shaecoing basis. Tenants ae dawn mainly fom the gou of fames ossessing some land on thei own Hatman and Boyce, 1983; Jansen, 1986; Rahman, 1986; Ahmed and Taslim, 1992. These obsevations have been exlained by landlods eluctance to lease out thei land to tenants with little to contibute othe than thei labo. and-owning tenants have a comaative advantage ove landless tenants in the acquisition of fam assets such as daught animals and iigation equiment, and they geneally have bette access to cedit. In addition, landless tenants may lack the necessay skills equied to cultivate the ented land oely e.g. choice of cos, oe land and wate management, selection and timely alication of inuts, etc.. This ae does not question the imotance of these factos. Instead it will be agued that inceased oductivity due to land owneshi is not the only eason why landlods efe landowning tenants. Because of its effect on noms and efeences, land owneshi makes the landed tenant moe attactive to the landlod by educing oblems of wok incentives. As will be shown, land owneshi contibutes to what Bowles et.al. 2001 have temed incentive enhancing efeences. and owneshi gives ise to individual taits that contibute to labo disciline, which ae valuable to landlods. The aim of this note is to show that once these consideations ae taken into account, we ae able to achieve a fulle undestanding of why landlods in Bangladesh have such a stong efeence fo landed athe than landless tenants. The next section esents some emiical evidence concening social oganization in ual Bangladesh. A simle model of the shaecoing elationshi is then fomulated, showing that when effot is non-contactible vaiations in efeences will affect the tenant s choice of effot and hence the income of the landlod. 2
2. Some Evidence fom Rual Bangladesh In most of the theoetical liteatue on agaian oganization, diffeent foms of land tenancy and labo contacts ae modeled togethe as substitutes along a continuous sectum of ossible contacts. 2 But this assumtion does not coesond so well to the actual social context of ual Bangladesh. Bangladesh ual society is highly diffeentiated, with ich landownes at the to and landless wage laboes at the bottom, while owne cultivatos and shaecoes come in between Jannuzi and Peach, 1980; Rahman, 1986; Jansen, 1986. 3 A eson with abundance of land can exect to be teated favoable by othe individuals, and this may give ise to both economic and social ewads though tansfe of maket good, maiage, tansfe of authoity, esectful teatment, etc.. Because of these ewads, individuals seek to maintain and incease thei social status though aoiate choice of actions and investments in land. At the othe end of the social ladde, smallholdes enjoy sueio status comaed to wage laboes. Thee is stong social stigma attached to being a wage laboe in ual Bangladesh. Woking as a wage laboe in the casual labo maket means being socially deogated to the bottom ung, o in the wods of Rahman 1986; "... ooe households with abundant labo sulies ty had to find agicultual land fo shaecoing. Thee ae social essues to emain in cultivation athe than become wage laboes in ual Bangladesh.... The shaecoes, though getting no ositive etuns fom cultivation afte deducting the cost of thei labo, still hang on to the system... as they deive some sychological satisfaction 2 An undelying assumtion in almost all the theoetical liteatue on contactual choice is that economic agents usue only thei own mateial self-inteest, and do not cae about social values and noms. Vey often, the theoy of the incial-agent elationshi seves as a majo tool to detemine otimum contacts that ae suosed to stike a balance between isk and incentive effects. See, e.g., Stiglitz 1988 and Hayami and Otsuka 1993 fo an oveview of the liteatue. 3 See also Scott 1976 who notes: The whole conventional hieachy of status among the ual oo is usually smallholde, tenant, wage laboue. These ae not, of couse, mutually exclusive categoies, since it is common to find cultivatos who simultaneously own some land and fam additional land as tenants, as well as wage laboues who have a lot of thei own. Yet, these categoies have had a social eality in efeence and status in the countyside desite the fact that the categoies could ovela consideably in tems of income.. 35. 3
fom being kishaks fames.". 163. Hence, smallholdes usually do not egad wage emloyment as an accetable altenative, at least in the shot un, as is usually assumed in the theoetical liteatue. At the same time, the geat majoity of smallholdes in Bangladesh have too little land elative to thei endowment of labo owe. They theefoe face the oblem of an excess of labo owe that cannot be gainfully emloyed on thei own land. A solution to this oblem may be to lease-in land. and leasing is theefoe attactive fo at least two easons. Cultivation of leased land ovides a solution to the oblem of emloying excess labo owe. Secondly, cultivation of leased land is a socially accetable solution. By being a shaecoe you kee you social status as an owne-cultivato. Howeve, having access to a shaecoing contact does not necessaily eesent a long-tem solution fo a tenant. In Bangladesh, most shae-contacts have a shot duation, usually one yea Januzzi and Peach, 1980; Singh, 1988. A shot-tem lease enables the landlod to evict the tenant if he is not satisfied with the tenant s efomance. 4 The theat of eviction can thus be seen as an endogenous enfocement mechanism that secues the landlod s objectives in the elationshi. 5 By aying the tenant an income slightly highe than the tenant s next best altenative, couled with a theat of eviction, the landlod can induce the tenant to wok had on the shaecoed land when effot is noncontactible. Fo a land-owning tenant, eviction does not simly deive him of the ootunity to ean an income fom the shaecoed land in the cuent eiod. He may also be foced to ente the casual labo maket in ode to meet his minimum subsistence need. Because of the social stigma attached to such 4 As noted aleady by Adam Smith in his Wealth of Nations, thee is a disincentive effect inheent in shae-contacts. Since a shaecoe only eceives a faction of his maginal oduct of labo, he has an incentive to ut in less effot on the land than he would if he owned the land himself. 5 See also Jansen 1986 who eots fom village Bangladesh that; Though inteviews with the olde men in the village we eceived the clea imession that shaecoing contacts lasted much longe a few decades ago. The consequences of the shotening of shaecoing contacts ae that moe lots fo shaecoing ae on the maket. [This] means less secuity fo the tenants and inceased ossibilities fo the well-to-do landownes to take advantage of the insecue situation.. 174-175. 4
wok, this eesents an additional social cost fo the landed tenant. In contast, a landless fame who has aleady woked as a wage laboe will geneally have no objections about taking u wage emloyment as an altenative to shaecoing. He has no social status to defend. A landed tenant is theefoe likely to wok hade in ode to avoid being evicted, all else being equal. Hence, a landlod will efe a landed athe than a landless tenant. In the next section the agument will be claified with the aid of a simle model. 3. A Model of the Shaecoing Relationshi Conside a non-cultivating landlod leasing out a given lot of land on a shaecoing basis. At havest time, the co X is divided between the tenant and the landlod, with the tenant eceiving y = X and the landlod eceiving π = 1X, whee 1 < < 0. The size of the co deends on the tenant s wok intensity and the cae he invests in the tasks involved in the cultivation ocess summaized in the vaiable, i.e. X = f with f ' > 0. 6 The lease is closed fo a eiod of at least one co season and is eeated infinitely so long as the landlod is satisfied with the tenant s efomance. If the tenant s efomance is not satisfactoy to the landlod, the contact is not enewed and the tenant is evicted. The tenant s efomance, howeve, is not easy to veify in a comlex and uncetain agicultual envionment, and hence cannot be detemined by contactual ageement. The landlod has to ely on an imefect measue of the tenant s wok intensity and monito the tenant andomly. The obability that the tenant s efomance is not found satisfactoy by the 6 This ae focuses attention on landlods selection of tenants given that shaecoing has been adoted. Thee exists a lage body of liteatue that attemts to identify conditions unde which shaecoing is sueio to altenative contactual aangements. See e.g. Hayami and Otsuka 1993 fo a suvey of the liteatue. Shaecoing may efom a vaiety of functions. It emits isk shaing when insuance makets fail, it can seve as a mechanism to sceen wokes of diffeent qualities, and it can coect maket imefections fo inuts othe than land. This ae abstacts fom all this by studying a situation whee the landlod has aleady decided to lease out the land on a shaecoing basis. 5
landlod deends on the amount of labo inut sulied by the tenant on the shaecoed land, i.e. = with ' < 0. Hence, is the obability of eviction, which is deceasing with the amount of labo sulied by the tenant. The decision stuctue is as follows. Fist the landlod chooses the shae to maximize his own income, then the tenant chooses how much labo to suly on the shaecoed land. 7 At the level of effot exected by the landlod, wok is subjectively costly fo the tenant to ovide. We assume that the utility of the tenant deends ositively on income and negatively on labo inut in the following way 1 U y, = f c whee c is the tenant s disutility of labo assumed fo simlicity to be constant. When choosing how much labo to suly on the shaecoed land, the tenant must conside both shot- and long-tem costs and benefits. Woking less had today means moe time is available to dedicate othe activities, but a highe obability of eviction and hence less income late. The value of having a shaecoing contact can thus be defined as the esent value of the tenant s futue income, taking into account the obability of eviction. If is the tenant s ate of time efeence, the esent value V of exected utility of stating out as a shaecoe is given by 2 V U y, 1 V = 1 Z whee Z is the tenant s fallback osition, i.e. the tenant s best available utility fom altenative activities. 8 As Hayami and Kawagoe 1993 note; In the village community eveyone is watching eveyone. Gossi about one s misconduct is ciculated by wod of mouth faste than any moden means of 7 Models of this tye ae analysed in Shaio and Stiglitz 1984 and Bowles and Gintis 1990, both studying the emloye-emloyee elationshi. 8 Fo simlicity we assume that income and disutility of effot ae both evaluated at the end of the eiod. 6
communication..167. In such an envionment, tenants who have been evicted fo oo efomance will soon become known and landlods will not hie them as shaecoes. The altenative activity fo both the landed and the landless tenant is thus casual labo emloyment. Z can then be inteeted as the lifetime utility of a tenant when he becomes a casual laboe. 9 Equation 2 then says that the tenant eceives U duing this eiod lus the esent value V if not evicted, that occus with obability 1, lus the esent value Z if evicted, that occus with obability. 10 The tenant chooses labo inut to maximize the esent value of exected utility, given in 2. Solving fo V in 2 and maximizing with esect to we get the following fist ode condition fo the tenant see aendix 3 f ' V Z ' = c The fist tem of 3 shows the tenant s diect gain fom highe income when labo inut on the shaecoed land is inceased his shae of the maginal oduct of labo, while the second tem shows the exected indiect gain aising fom inceased labo inut. ' indicates the decease in the obability of eviction by inceasing labo inut, and V Z is the value of holding a shaecoing contact the diffeence between the value of utility fom being a shaecoe and the value of utility fom being a casual laboe. The quantity V Z can thus be inteeted as the cost of losing tenancy, a cost which the tenant seeks to avoid though had wok. 11 Accodingly, equation 3 says that, fo any given shae, the tenant will detemine how had to wok on the shaecoed land by tading-off the diect gain fom inceased labo inut lus the effect that 9 Recall that the tyical landed tenant has too little land to secue an income ove the subsistence minimum. He is theefoe deendent on altenative activities to emloy his excess labo owe. 10 This fomulation imlies that we conside the landlod-tenant elationshi as an infinitely eeated game. 7
additional effot has on the obability of etaining the contact, against the maginal disutility of effot. Equation 3 defines the tenant s eaction function, which shows the level of effot by the tenant fo any given shae. This effot esonse function is inceasing in see aendix as figue 1 shows. Figue 1. Equilibium shae and effot levels fo the landlod and the effect of a eduction in the tenant s fallback osition. Shae 1 2 3 π 2 π 1 1 π 2 1 2 Effot 11 Clealy, the landlod s theat of eviction is cedible only if V > Z. If V = Z, the tenant is indiffeent between being a shaecoe and altenative emloyment ootunities and it makes no sense fo the landlod to suevise the tenant s labo inut. The tenant would not cae about the suevision, since he is in a osition to shift to othe activities at no cost. 8
The landlod detemines the value of the shae as to maximize his own income π = 1 f, knowing that a highe shae fo the tenant induce him to ut in moe effot. 12 The esulting equilibium ai 1, 1 is illustated in figue 1. The iso-income cuve π 2 eesents shae/effot combinations yielding the same income fo the landlod. Pefeed iso-income cuves lies to 1 2 3 the ight, i.e. π > π > π. Both the landed and landless tenant will choose labo inut so as to satisfy the fist ode condition, given by 3. Gahically this means that they will choose an effot level along thei eaction function. Howeve, the stuctue of social elations in which the landed tenant is embedded enhances his wok incentives comaed to the landless tenant. If evicted, the landed tenant may be foced to ente the casual labo maket, which means being socially deogated. In contast, the landless tenant has no social status to defend. If evicted, he loses the contact ent, but has no objection taking u waged emloyment as an altenative to shaecoing. The quest fo social status will, in othe wods, incease the desiability of etaining the contact fo the landed tenant. If the landed tenant can secue a enewal of the lease, he can make a fulle utilization of his othewise unemloyable labo owe. If the lease is teminated, at of his labo owe will emain unde-utilized, altenatively he may be foced to ente the casual labo maket and be socially deogated. The steam of utility fo the landed tenant in altenative activities is thus lowe comaed to the landless tenant. This can be modeled as a eduction in Z fo the landed tenant, aising the maginal subjective benefit of effot. A eduction 12 When making an offe to the tenant the landlod has to balance two effects woking in oosite diections. An incease in the shae will enhance the tenant s effot, aising outut on the shaecoed land and thus the landlod s income. But at the same time an incease in the shae also educes the landlod s income. As long as the indiect gains fom an inceased shae exceed the diect loss ion income of the incease, a highe shae will be offeed and the cost of loosing tenancy will be ositive. 9
in Z shifts the tenant s eaction function downwads see aendix, making the land-owning tenant moe ofitable to hie. Inceased income fo the landlod follows diectly fom the fact that he gets moe effot fom the tenant fo each shae offe. Figue 1 shows the effect of a decease in Z. The new equilibium ai 1 2, 2 is illustated by the tangency oint between the iso-income cuve π of the landlod and the new effot esonse cuve, leading to a eduction in the tenant s shae and an inceased effot. Clealy this is in the inteest of the landlod. Hence, the landlod will get a highe etun if he leases out the land to a landed athe than a landless tenant. 4. Concluding Remaks The main objective of this ae has been to exlain why landlods in ual Bangladesh have such a stong efeence fo landed athe than landless tenants. The existing liteatue has emhasized that fames with some land of thei own ossess fam assets and skills that make them moe oductive as tenants. But thee is also be anothe mechanism that may exlain why landowning fames ae attactive as tenants. The hieachy of social status in ual Bangladesh tansfoms land-owning tenants efeences in ways that ae ofitable fo landlods. The geate cost of losing tenancy fo a landed tenant, stemming fom his socio-sychological avesion to offeing his labo in the casual labo maket, induces him to cultivate the ented land with geate cae and intensity comaed to a landless tenant with no social status to defend. Theefoe, the landlod gets a highe etun if he ents out his land to a landed athe than a landless fame. 10
Aendix Solving fo V in 2 gives A1 V f c Z = Maximizing A1 with esect to gives A2 f ' c ' Z f c Z ' 2 = 0 f ' c ' Z V' = 0 which gives the fist ode condition 3. Poof that is inceasing in : It will be assumed that the second ode condition fo a local maximum is fulfilled. Hence, D defined by A3 must be negative A3 D = f '' V Z '' < 0 Diffeentiating 4 totally with esect to, we get dl d A4 f ' f '' V ' ' V'' = 0 whee d d 11
12 ' ' ' f d d Z V c f f V = = Inseting fo ' V in A4 we get A5 0 ' ' 1 > = f f D d d Poof that an incease in Z shifts downwads the tenant s eaction function: Diffeentiating 4 totally with esect to Z, we get A6 0 '' ' '' ' ' ' ' = dz d Z dz d V V dz d f Z whee ' ' ' dz d Z V c f V Z = = Inseting fo Z V ' in A6 we get A7 0 ' 1 < = D dz d An incease in Z lowes, fo a given shae, and vice vesa.
Refeences Ahmed, F.U. and Taslim, M.A. 1992. An Analysis of and easing in Bangladesh Agicultue, Economic Develoment and Cultual Change, 615-628. Badhan, P. ed. 1989. The Economic Theoy of Agaian Institutions. Oxfod. Claendon Pess. Bowles, S. and Gintis, H. 1990. Contested Exchange. Politics & Society, 18, 165-222. Bowles, S., Gintis, H., and Osbone, M. 2001. Incentive-Enhancing Pefeences: Pesonality, Behavio and Eanings. Ameican Economic Review, 91, 155-158. Hatman, B. and Boyce, J.K. 1983. A Quiet Violence; View fom a Bangladesh Village. ondon. Zed Pess. Hayami, Y. and Kawagoe, T. 1993. The Agaian Oigins of Commece and Industy A Study of Peasant Maketing in Indonesia. ondon. Macmillian Pess. Hayami, Y. and Otsuka, K. 1993. The Economics of Contact Choice; An Agaian Pesective. Oxfod. Claendon Pess. Jannuzi, T.F. and Peach, J.T. 1980. The Agaian Stuctue of Bangladesh; An Imediment to Develoment. Coloado. Westview Pess. Jansen, E. 1986. Rual Bangladesh; Cometition fo Scace Resouces. Oslo. 13
Nowegian Univesity Pess. Rahman, A. 1986. Peasants and Classes; A study in Diffeentiation in Bangladesh. ondon. Zed Pess. Scott, J.C. 1976. The Moal Economy of the Peasant. New Haven. Yale Univesity Pess Shaio, C. and J. Stiglitz, J. 1984. Equilibium Unemloyment as an Woke Disciline Device, Ameican Economic Review, 74, 433-444. Stiglitz, J. 1988. Economic Oganization, Infomation, and Develoment. In Cheney, H. and Sinivasan, T.N. Handbook of Develoment Economics, Vol. I. Amstedam. Noth Holland. 14