:a 'D= r- w- - ; - - - - : - - _ 1 7 C 4 1 ( o pgjs, >- 7-W Fsz ~N==~-=. ______S:: A joint publication by the Development Research Group and the Poverty Reducton Group of the World Bank No. 9 November 1997 Public Works Programs: - e-dsge Whlat are the Income Gains pic works to the Poor? pr c .s..elf-argeted to'i the woub kc o programs have been a Opportunity Costs p o a can 1 popular policy i nt for pov- t a =eny aleviaio in developing Wa cvt ospbcwrsen -v e-... .zo. .- cournes. What do the poor gain mn tm p e d e? If it s p n n gains of icorm fm such progams? A World unemployment then te opportunity Bank study takes a fresh look at ths issue (forgone nome) of participaton in scha- examining how households respond to scheme will be regligible. However, pub- - - -p employmrt pporunities provided by lc works employment may displace public w ois p 1i, woikers fiom rher income-earing oppor- Thestudy draws onhouseholddatafor uites, such as altemative sources of reallocated from own-nim activities. The the villages of Shirapur and Kanzara i the wage labor or own-frm acvites he ovelall level of EGS employment was also state of MAaharashtra, ndia, where publc eanings that would have been associated lower i KanZra, consise wit rela- works employment is available under the hes,e displaced actvites represent e tively higher average mcomie levels in the state's "eniployment Guarlanee Scheme Pecuiw opportuity cost of publc village; however, given the patten of dis- (EGS). An explicit aim of this scheme is works employment. is the at3 ated ogpportunity cost to alleviate rural povertv through income The results of the study suggest that he her, at 32ofgross EGS eaing gains to all tcipateig workers. The EGS C7oppo°ity costs to participants in the The assessment of t ensfar benfits offers portunities for unkiers nTaeuaG EGS are geeay low, even though re- yields marnedly different, and misleading, suits differ somweihat by Ollage. r p market wage iS labor at low wages (on a par with the Shirapur, a relatively poor vi used to represenit the opportunity cost of The data for the villages are taken from sghtly more than 40% of the bmes public wors employment. For istance, a longitudinal household survey conducted by laborms on the EGS site replaced lei- agen ho u ral age yes arer- by well-trained, resident investigators for sure and dmestic work, while one-third the International Crops Research e replaced unemploymert. Only one-fifth estimates of foregone incomes are 93% of of the Semi-Arid Tropics (ICRISTAT). involved a sacrifice of other wage labor gross eamnigs from public works m The data are widely considered to be of time The value of foregone incomnes is Sbirapur, and 77% in Kanzara. This ex- high quality, providing a unique opportu- esfimated to be 21% of gross wage a- ercise ilustgates tha while the technte nitv to examine household behavior i ings from EGS employment inShirapur of usig the prevailig market wage to Tesponse to the Maharashtate emn- The pattem is rather differen in Kan- represet oppouity ost is relatively ploymetnl scheme over a six-year peiod zara, where nearly a third of time on the smple, it clearly overestimates the fore- (1979-80 to 1984-85). public worksite displaced other wage labor gone icomes of workers participatng m activities, and one-quarter represented tim the public works program. Do Public Works The study points to a number of addi- If public works primaily displace un- Reduce Poverty? tional factors that remain to be considered employment, foregone eamungs wll be In a typical year, net twansfer beefits before a proper assssment of the: per- negligible. However, displacement of from public works employment generated formance of public wvorks progams other income-eing activities will a 10% increase in preransfer earnings relative to alternative policies with the represent a pecuniary opportunity cost to among households participating in the same aim can be achieved. For example, participation in the scheme, thereby reduc- EGS m Shirapur, and a 7% increase in the total impact of the EGS on poverty is ing the net income benefits of the Kzarza These net increases in eamings probably undriated, since any in- program. unambiguously reduced poverty, for ay duced iincome effects of public wvorks The evidence from Idia suggests that poverty line or measure within a broad projects (such as through asset creation, the pecuniary opportnity cost to public class. and the impact on agricultural wages) are works participants is, in fact, low. This Iterestgly, a simple coun rfituctal test not factored in to the present calculations. result emeiges most clearly in the rela- indicates that a uniform transfer of the Yet the results presented here do make tively poor village of Shirapur, where same gross amount across all households clear that in assessing policy effetiveness, participation in the program is also more in both villages (whether poor or not) it is important to control for the costs of widespread. would achieve similar reductions in the targetng through public works progrmms. Overall, the public works program in proportion of the population in poverty- These costs-namely foregone incomes India does appear to generate sizable net regarcless of where the actual poverty line and non-wage expenditures-must be income gains to participts, rangg from is drawn-and mnay acbieve greater im- taken into accmt when calculating post- 7% to 10% in the villages studied. These provenents in the equality of income tansfr income gains to poor households tansfer benefits also lead to an unambi distribution among the poor (reductions in rug from public works employment. guous redion in poverty measures. the severity of poverty). Nonetheless, the results from such a test Summary do not imply that the poor would be A recet World Bank study draw on clearly better off if rural publc works dt f t vlg to preset a projects m Inda were abandoned m favor n e .> ^ . , . ~~~~new empircal approach to assessiq;g the of uniform ransfers. The simulation re- imact of publc works sult assme egligile aminsS tpac ofpublc wlksprograms on net suts assme neglible administrative household income and poverty. The study costs of unifom transfers, along with a estimates the opportuniy cost of partici- negligible inpact on the pre-intervention in public . . . pabon m pupationworks enipk,yment distribution, neither assumrption is relis- programs by asking the question "what Sue Gaumav Datt and Martn RavafliorL 1994. tiC, and both probably lead to an activity is public works employment dis- 'ww Gow for the Poor firm Publc Works overesmation of the impact of uniform placnngpl omn LS Wolding Paps No. 100. tansfers on poverty. Ihe World Bankc Washg DC. 49 pp. PovertLines is prepeteH by Galee Hall w fth assistance fom fim Safer. This issue was edited by Bnan iCulbane. For information orfor a subscniption, contact: LSMS Administat (Antn: Poverty Lines) Isit our World Wide Web sites: Word Bank, N8"58 Poverty Group Home Page: hbpJ/hcworldbanlorg4nl hcov'poprmty/ccntes.htfr 1818 H Stree N.W. LSMS Home Page: httpi/hv.worldbankorg Iprdlphsinssn/lsmhorne.hti Washington, D.C 20433 E-mnail address: Ismsaworldbankorg Fax: (202) 522-1153