* $i Policy, Research, and External Affairs WORKING PAPERS Development Economics The World Bank October 1991 WPS 789 Intrahousehold Inequality and the Theory of Targeting Lawrence Haddad ati lu Ravi Kanbur Hlere is a start at l1nking the literatures on taroetino anzd on illtraho0l.lehold IneIualiy lsZ V h iih hav e developed fpidlv hutll largely independent of each other. . Lho(x..r4$-r :c L, ;F I . JE:c8! I;J., 4U 4 *. *'. : -. r dI\c. :N _ : .-'v : .. . ] Lhad _' Os,-L.C.:tJ' 'wtr> . II' '" ..I.: *x' ...Qdt.i :.~.C.Cc . ( . '.:2'F I .. *2S77...~ ..: (.:..r-' 4'¾.": . '. ..z V.. V .- Policy, Research, and External Affairs Development Economics WPS 789 'I'llis papker It product ol' die Research Advisory Staff', Oilice of Ilhc Vice Prcsidcit, Development lcnnormi%s - is part ot a largcr cl'lort in PRE to understand the design of poverty alleviation policies. Copies are availablM frce fIrom the XVorl'J Biank 1818 If Street NW WaW hington D)C 20433. Please coiltact Janc SweenC\, rk,om S3-026, extension 310(21 (13 pages). October 1991. Th,c te.o 1 ltCatiufcs oll la!tlgt inand on intrahiouseihold inequality anid allocation mecha- initrahousehold incqualit\ liavc developcd rapidly nisms could affcet the standard analysis of' oVkc, the pa> 13 I * cars, but la Rcl\ independent of targeting theory. Thex coniclude with a list of each other. policy questions for further rcsearcih including the following: T'hc literature on waroetino concerns itself' ; ilt)h Il desi Ln of t.1a dld tranlslfcr puitgi slwl * M iOW are coiincut ional rules lor indicator pmo eny alleviation in the presence of'limitcd targeting modif'ied by different housciold in tlormiation on l lho the poor are. allocatioIn mnechanisms. 'I'lTh literaturc oii intrahouschold inequ al ity * How far v" roing can onlC go in targetinlg by arosc out ot a dissailst clion wnitih "unilar'' simpl) assumiing that int-ahouseihold inequality modf,ls otheiC houschol(j. es.plecially in explain- does n0ot exist, when in fact it does?' ill_ ilkqUlid t in consumtipion and achievements of (difteellt household mriembcrs, even al;ter * What sort of intrahouscihold inlonnation allou. i r tor reie\ adi illerences a.imolln tleilm. s,tould bc collected to best aid tareeting? Iladdad and Kanhur hbein io tforge ihe link * f-lo% do the "bargaining" versus the "com- hetu.cen the tvo literattu-es, so thex canl address mon prefcrence" views of tfhe liousciold irillu- issues policxmakeurs face around iltle u.old, encc our evaluation of alternative transfer Alter a hrief reprise of ilic key featurcs of thc prograrms? t\ko litclrture.u tilc indicale ho% tile pilesence ol' I11<1[iSl \\, 'i,} . I' 1 l'l S, t ;. :. ":.,.. 1:t' tl'i2u:1,wl>. c 1 kertk un,Lr y '\ '! I iI3. tld1k '. jlii . Rcxr.tr 'e (in~t7 f\WflUi, 'c p'llvi, tit no: flCk c-.I!i\ o1h c''n j' ii H.,i~k poli Pict .:- .! 1,% tl:, I'R L Di;senn;:m.,r-. Cc fitcr INTRAHOUSEHOLD INEQUALITY AND THE THEORY OF TARGETING* CONTENTS 1. INTRODUCTION ...........1. 2. INTRAHOUSEHOLD ALLOCATION AND INEQUALITY . . . . . . . . . . . 2 3. THE PRINCIPLES OF TARGETING . . . . . . . . . . . . . . . . . 4 4. INDICATOR TARGETING AND INTRAHOUSEHOLD ALLOCATION . . . . . . 6 5. SELF-TARGETING MECHANISMS AND INTRAHOUSEHOLD BARGAINiING . . . 8 6. FURTHER RESEARCH . . . . . . . . . . . . . . . . . . . . . . . 10 * Invited paper for presentation at the Sixth Annual Congress of the European Economic Association, Cambridge, England, September 1991. INTRAHOUSEHOLD INEQUALITY AND THE THEORY OF TARGETING 1. INTRODUCTION The two literatures on targeting and on intrahousehold inequality have developed rapidly over the past 15 years, but largely independently of each other. The literature on targeting goes back at least as far as Akerlof (1979), and has concerned itself with the design of tax and transfer programs for poverty alleviation in the presence of limited information on who the poor are. The origins of the literature on intrahousehold inequality are referred to in Sen (1984). This literature arose out of a dissatisfaction with "unitary" models of the household, especially in explaining observed inequality in consumption and achievements of different household members, even after making an allowance for relevant differences among them. While the two academic literatures have indeed been mostly innocent of each other, the significance of intrahousehold inequality and allocation for targeting has not escaped the notice of policymakers. In developed countries, the debate on whether child benefit should be paid to the mother at a welfare office, or to the father's paycheck through a tax exemption, has essentially been about the perceived allocation, and reallocation, of intrahousehold resources. Similarly, the discussion in developing countries about knock on effects of supplementary feeding programs for children, and -2- for pregnant and lactating women, is essentially one about whether or not extra calories at the feeding station means fewer calories received by the individual at home. The object of this paper is to begin the task of forging a link between the two analytical literatures on targeting and on intrahousehold inequality, so that they are better able to address issues faced in this area by policymakers in developed and developing countries. After a brief reprise of the key features of the two literatures in Sections 2 and 3, Sections 4 and 5 will indicate how the presence of intrahousehold inequality and allocation mechanisms could affect the standard analysis of targeting theory, Section 6 concludes with a list of topics for further research. 2. INTRAHOUSEHOLD ALLOCATION AND INEQUALITY Are similar individuals treated dissimilarly in the allocation of consumption within a household? The answer from the empirical literature is that this possibility has to be taken seriously. Sen (1984) summarizes a number of studies on outcome variables which argue that girls within households are discriminated relative to boys. Harriss (1986) presents at least some evidence of calorie intake inequality. Haddad and Kanbur (1990a) shew that standard measures of inequality in calorie adequacy would be understated by 30 to 40 percent if intrahousehold inequality was ignored. But is the above sufficient evidence that similar individuals are being treated dissimilarly? If individuals differ in their -3- produdltiviiies, Lieri exArd caiories may weili De callea tor as an income generation device for the household as a whole (Pitt, Rosenzweig, and Hassan 1990). Even a household that was egalitarian in terms of its welfare functicn might allocate calories unequally. While there are some tests of the degree of inequality aversion displayed by a household welfare function, as revealed by consumption allocation and other outcomes (Behrman and Deolalikar 1989), a major debate in the literature is on whether such a "common preference" model is valid. An alternative is a view of the outcomes as being determined by intrahousehold bargaining. The empirical implications of this for demand theory are laid out by McElroy (1990), who makes clear that the difference between the common preference and bargaining approaches can be traced to the fact that, in the latter, changes in the external environment can alter the "threat points" and lead to reallocations that are different from those predicted by the household wel-are maximizing model. Haddad and Kanbur (1990b) have considered the predictions of the bargaining model on intrahousehold inequality as the household gets better ott. they find that the net effect depends on a subtle interaction between changes in the gains from cooperation and in the threat points. Under certain conditions, intrahousehold inequality can first increase and then decrease, thus tracing out a "Kuznets curve." Haddad and Kanbur (1990c) find empirical evidence for such a relationship for data from the Philippines. Thomas (1990) and Altonji, Hayashi, and Kotlikoff (1990) are among the authors who find -4- evidence that is not supportive of the common preference model. As we shall see, this finding is of some significance for targeting of poverty alleviation programs. 3. THE PRINCIPLES OF TARGETING Besley and Kanbur (1988) have considered the principles of targeting that underlie the literature emanating from Akerlof (1978). The theory of targeting concerns itself with the design of transfer mechanisms for alleviating poverty. Given a pre-intervention distribution of income and a poverty line, an ideal solution nmight be characterized as being one where sufficient income is transferred to the poor to bring them just up to the poverty line. There are, however, at least two problems with the ideal solution. First, since transfers fall one-to-one with income, it entails an effective marginal tax rate on the poor of 100 percent. This is bound to have incentive effects-Kanbur, Keen, and Tuomala (1991) give a quantitative feel for the consequences. Second, it requires that the policymaker has quite detailed information on individuals, making it possible to exclude those above the poverty line, and to ta,lor the magnitude of the transfer quite finely to those below the poverty line. Such detailed information, and the administrative ability to use it, is simply not present in most developing countries, and, in certain contexts, in developed countries as well. In view of this second problem, Besley and Kanbur (1988) characterize two types of targeting mechanisms-indicator targeting -5- apd self-targeting. Indicator targeting relies on making the transfer contingent not on income or consumption, but some easily observable characteristic, such as sex, age, landholding, region of residence, etc. Akerlof (1978) referred to this as "tagging." As he shows, this extra information is bound to be useful. Kanbur (1987) and Kanbur and Keen (1989) derive optimal rules for contingent transfers, while Ravallion (1989) quantifies the gain from using this information. Haddad and Kanbur (1991a) develop the theory of "upper-limit indicator targeting," where an upper cutoff of an observable variable, such as age, is used to determine eligibility in transfer programs (e.g., supplementary feeding). Self-targeting mechanisms rely on an announced scheme that permits unlimited participation, but is designed in such a way that only members of the target group find it worthwhile to participate. Thus, costly administrative screening and verification of indicators, etc., is not necessary-the incentives to participate are themselves the screen. Such "self-screening" or "self-selection" has been analyzed in the recent theoretical literature on imperfect information (e.g., Hoff and Stiglitz 1990), but the mechanisms themselves were well known, for example, to those who designed the Indian famine codes in the 19th century (Dr6ze 1988). The basic idea is to impose a cost of participation that varies directly with pre-intervention income. The best known illustration of such a scheme is the Employment Guarantee Scheme (EGS) in tne state of Maharashtra in India (Ravallion 1991). This scheme, and others, rely -6- on ofi¢;10 ni un. -I m..tcj n 'LI , but at d Waye low enough to attract only those with low pre-intervention incomes. Ravallion (1989) presents a theoretical analysis of such schemes, while Ravallion (1991) surveys some of the empirical evidence, concluding that such schemes do, indeed, act as effective targeting devices. But Ravallion's analysis, and most of the other literature on self- targeting schemes, ignores the effect that they can have on intrahousehold allocation, as we shall see in Sectiot, 5. 4. INDICATOR TARGETING AND INTRAHOUSEHOLD ALLOCATION The simple analysis of indicator targeting, as conducted, for example, in Kanbur (1987), is in terms of lump sum transfers to individuals with different values of an indicator variable (e.g., region or landholding). All individuals with the same indicator value are treated identically, since, ex hypothesi, there is no other information to differentiate between them. Un this assumption, Kanbur (1987) derives leakage minimizing rules based on statistical information derived, for example, from sample surveys. Individual incentive effects are ignored in the simple analysis. Kanbur and Keen (1990) show how the rules are modified when labor supply is elastic, so that incentive effects matter and the intervention itself changes the distribution of pre-intervention income. Consider now the assumptions underlying indicator targeting in light of the literature on intrahousehold allocation. Suppose the indicator variable is such that not all members of the household have -7- the same indicator value, e.g., when a program screens by age or by gender. The assumption of lump sum transfers to individuals is then seen to be making an assumption about the nature of intrahouseiold allocation. Is it that the individual who receives the transfer gets to keep it? How might we rationalize this? Certainly not in terms of a "common preference" model of the household, one of whose implications is that all household resources are first aggregated into a common pool and then reallocated, so that the marginal recipient of the transfer will, in the final analysis, only benefit partially. Such an argument is valid even if individual-level incentive effects are taken into account. This is an incentive effect of a different sort, to do with reallocations among groupings of individuals who have different values of the indicator variable being used. Suppose now that the indicator variable is such that, by definition, all members of a household must have the same value, e.g., region of residence or crop/tenure group. What does the assumption of equal lump sum transfer now mean? It means that every member of the household receives the same transfer. But how is this effected? The slightest reflection on design and implementation issues will reveal that this assumption cannot possibly be maintained in terms of administrators actually ensuring an equal transfer to each member of a household. And, in any case, there is the issue of intrahousehold reallocation after the equal transfers take place. If the transfer in practice is effected through the head of household, say, then the -8- assumption of equal transfer to all members of the household implies a model of introhousehold allocation which may or m-y not be justified. In any event, what is clear is that the analysis of indicator targeting cannot proceed without a well articulated view of how a household allocates resources to its members. As Haddad and Kanbur (1991a) show, whether or not intrahousehold reallocations take place-and their exact nature--can make a big difference to who ends up benefitting from a program that relies on indicators to target transfers. 5. SELF-TARGETING MECHANISMS AND INTRAHOUSEHOLD BARGAINING Haddad and Kanbur (1990b) present a simple model of intrahousehold bargaining whose object is to relite the gains from cooopration, and the outside options, to agricultural production possibilities. In that model, there are two individuals, each with access to a production function that produces output as the result of two task inputs. There is comparative advantage in the tasks, so it pays to cooperate and specialize in tasks. But how are the gains from cooperation to be divided? Suppose that the fallback option for each individual is identified with the outcome of working alone. The individua. with absolute advantage thus has a better fallback option and, if we interpret this as a threat point of a Nash bargain, he will get the higher share of the cooperative output. There is, therefore, intrahousehold inequality. -9- Haddad and Kanbur (1990b,c) trace out the implications of this model as productivity of the two individuals increases so that the household as a whole is better off. But what does this model imply for the effectiveness of self-targeting mechanisms, such as the Employment Guarantee Scheme? Haddad and Kanbur (1991b) present a simple analysis which is nevertheless illuminating. Consider the case where total cooperative output is giveln by x and the allocation of each individual is x1 and x2. Let the threat points be S, and S., defined by output when the two individuals work on their own. Suppose that S, S,, so that, under the axioms of Nash bargaining, x1 > x2. Now, suppose that the government introduces a scheme that guarantees an income of W. How will this affect intrahousehold inequality, and, in particular, the well-being of the less well off individual? For W ¸ S,, of course, there is no effect whatsoever. But when W > S2, W becomes the threat point of individual fl *.J-..1 6LL- *..**..1 OL ...Ji-le Ii L b1 rY jjj9 L. UIlUCi tlIIC UsuaI uutI u 1e Of Ns Vh4iIIiy, Lnererore, if W < x2, then an increase in W will increase x2 and decrease x1, even though the guarantee is not taken ug. As W increases to S,, both threat pu;iiLb tLeI.uUi e qu!' a-,u tihere is perfect equaiity. ror w > 4x, bOth individuals will move to the guarantee scheme. What is remarkable is that, even before this point, the scheme has a long reach-it equalizes intrahousehold allocation by altering outside options. Of course, the credibility of the guarantee is at the heart of the matter, and this brings the issue back to some of the policy debates on employment rationing of employment guarantee schemes -10- (Ravallion 1990). With rationing, the guarantee is not credible, and will not have its full effect on intrahousehold allocation. The analysis of interactions between intrahousehold allocation and self- targeting schemes thus stands cut as a potentiallh fruitful topic for theoretical and empirical research. A start is made in this direction in Haddad and Kanbur (1991b). 6. FURTHER RESEARCH We hope that the juxtaposition of the two literatures that is attempted here will serve to increase the interaction between them. Such cross-fertilization has already begun, but a host of interesting theoretical, empirical, and policy questions suggest themselves. Among these: How are conventional rules for indicator targeting modified with different intrahousehold allocation mechanisms? How far wrong can one go in targeting by simply assuming that intrahousehold inequality does not exist, when, in fact, it does? What sort of intrahousehold information should be collected to best aid targeting? How do the "bargaining" versus "common preference" views of the household influence our evaluation of alternative transfer programs? In particular, what is the quantitative significance of the "long reach" of guarantee and self-targeting mechanisms, through their effect on bargaining threat points, even when no household member actually participates in the scheme? -11- REFERENCES Akerlof, G. 1978. The economics of 'tagging' as applied to the optimal income tax, welfare programs, and manpower planning. American Economic Review. Altonji, J. G., F. Hayashi, and L. J. Kotlikoff. 1989. Is the extended family altruistically linked? Direct tests using micro data. National Bureau of Economic Research, Cambridge, Massachusetts. Mimeo. Behrman, J., and A. Deolalikar. 1989. Health and nutrition. In Handbook of development economics, ed. H. Chenery and T. N. Srinivasan. North Holland: Amsterdam. Besley, T. J., and R. Kanbur. 1988. The principles of targeting. Forthcoming in Including the poor, ed. M. Lipton and J. vander Gaag. Washington, D.C., U.S.A.: The World Bank. Dreze, J. 1988. Famine prevention in India. Discussion Paper No. 3. Development Economics Research Proaramme, London School of Economics. Haddad, L., and R. Kanbur. 1990a. 1-'sw serious is the neglect of intrahousehold inequality? Economic Journal. . 1990b. Are better off households more unequal or less unequal? PRE Working Paper No. 373. Washington, D.C., U.S.A.: The World Bank. -12- __* .1990c. Is there an intrahousehold Kuznets Curve? PRE Working Paper No. 466. Washington, D.C., U.S.A.: The World Bank. . 1991a. Upper-limit indicator targeting and age-based nutritional interventions: Optimality, information, and leakage. Discussion Paper No. 107. Development Economics Research Centre, University of Warwick. _ 1991b. Public employment schemes and intrahousehold inequality. Forthcoming. Harriss, B. 1990. The intrafamily distribution of hunger in South Asia. In The political economy of hunger, Volume I: Entitlement and well-being, ed. J. Dreze and A. K. Sen. Oxford University Press. Hoff, K., and J. E. Stiglitz. 1990. Imperfect information and rural credit markets-Puzzles and policy perspectives. World Bank Economic Review. Kanbur, R. 1987. Measurement and alleviation of poverty. I.M.F. Staff Papers. Kanbur, R., and M. Keen. 1989. Poverty, incentives, and linear income taxation. In The economics of social security, ed. A. Dilnot and I. Walker. Oxford University Press. Kanbur, R., M. Keen, and M. Tuomala. 1991. Optimal nonlinear income taxation for the alleviation of income poverty. PRE Working Paper No. 616. Washington, D.C., U.S.A.: The World Bank. -13- McElrcy, pA l90Q. Th-) . of Nah-bargained how hofd behavior. Journal of Human Resources. Pitt, M., M. Rosenzweig, and M. Hassan. 1990. Productivity, health, and inequality in the intrahousehold distribution of food in low- income countries. American Economic Review 80 (5): 1139-1156. Ravallion, M. 1989. Land-contingent poverty alleviation schemes. World Development. _ 1990. On the coverage of public employment schemes for poverty alleviation. Journal of Development Economics. . 1991. Reaching the rural poor through public employment: Arguments, evidence, and lessons from South Asia. World Bank Research Observer. Sen, A. K. 1984. Family and food: Sex bias in poverty. In Resources, va7ues, and development, ed. A. K. Sen. Oxford: Blackwell. Thuzas, L. a199. I,trahAusehoid r uurce allocation: An inferential approach. Journal of Human Resources. PRIE-Work no Paper Series Contact Title AutAor Date for paer WPS771 Macroeconomic Structure and Ibrahim A. Elbadawi September 1991 S. Jonnakuty Policy in Zimbabwe: Analysis and Klaus Schmidt-Hebbel 39074 Empirical Model (1965-88) WPS772 Macroeconomic Adjustment to Oil Ibrahim A. Elbadaw' September 1991 S. Jonnakuty Shocks and Fiscal Reform. Klaus Schmidt-Hebbol 39074 Simulations for Zimbabwe, 1988-95 WPS773 Are Gnana's Roads Paying Their Reuben Gronau September 1991 J Francis Way9 Assessing Road Use Cost and 35205 User Charges in Ghana WPS774 Agricultural Pricing Systems and Mark Gersovitz October 1991 B. Gregory Transportation Policy in Africa 33744 WPS775 The Macroeconomics of Public William Easterly October 1991 R. Martin Sector Deficits A Synthesis Klaus Schmidt-Hebbel 39065 WPS776 Enforcement of Canadian J'Jnfair" Mark A. Dutz October 1991 N Artis Trade Laws The Case for Competition 37947 Policies as an Antidote for Protection WPS777 Do the Benetits of FiAed Exchange Shantayanan Devarajan October 1991 A. Bhalla Rates Outweigh Their Costs9 The Dani Rodrik 37699 Franc Zone in Africa WPS778 A Dynam!c Bargaining Model of Eduardo Fernandez-Arias October 1991 S. King-Watson Sovereign Debt 31047 WPS779 Special Programme of Research, Janet Nassim October 1991 0. Nadora Development and Research Training 31019 in Human Reproduction V415/00 Optitfidi User unarges ana Cost Ian G. Heggie October 1991 P. Cook Recovery for Roads in Developing Vincy Fon 35462 Countries WPS781 The Korean Consumer Electronics Taeho Bark October 1991 N. Artis Industry. Reaction to Antidumping 37947 Actions WPS782 The Economic Effects of Widespread Patrick Conway October 1991 N. Artis Application of Antioumping Duties Sumana Dhar 37947 to Import Pricing WPS783 The Origins and Evolution of J Michael Finger October 1991 N Artis Antidumping Regulation 37947 WPS784 Chemicals from Poland. A Tempest Andrzej Olechowski October 1991 N. Artis in a Teacup 37947 PRE Working Papei Series Contact Tole Atnbhr Dat for 12Ver WPS785 How Did the Asian Countries Avoid lshrat Husain October 1991 S. King-Watson the Debt Crisis? 31047 WPS786 Fiscal Policy for Managing Sadiq Ahmed October 1991 B. Prasertwaree Indonesia's Environment 82477 WPS787 Private Investment Under Macroeco Klaus Schmidt-Hebbel October 1991 S. Jonnakuty nomic Adjustment in Morocco Tobias Muller 39074 WPS788 How Expectations Affect Reform Francesco Daveri October 1991 S. Jonnakuty Dynamics in Developing Countries 39074 WPS789 Intrahousehold Inequality and the Lawrence Haddad October 1991 J. Sweeney Theory of Targeting Ravi Kanbur 31021