THE WORLD BANK '45)~~~~~~~~~~~~~~~~~~~~~~4 >7 .s rg 2 r iUt 2* August 2 Other Titles in This Series Dynamic Risk Management and the Poor-Developing a Social Protection Strategy for Africa Engaging with Adults-The Case for Increased Support to Adult Basic Education in Sub-Saharan Africa Inclure les adultes - Pour un appui a l'6ducation de base des adultes en Afrique subsaharienne Enhancing Human Development in the HIPC/PRSP Context-Progress in the Africa Region during 2000 Early Childhood Development in Africa-Can We Do More for Less? A Look at the Impact and Implications of Preschools in Cape Verde and Guinea Le d6veloppement de la petite enfance en Afrique-Impact et implications des centres pr6scolaires a Cap-Vert et en Guin6e AIDS, Poverty Reduction and Debt Relief-A Toolkit for Mainstreaming HIV/AIDS Programs into Development Instruments Africa Region Human Development Working Paper Series Systemic Shocks and Social Protection Role and Effectiveness of Public Works Programs K. Subbarao Africa Region The World Bank ii AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES C August 2001 Human Development Sector Africa Region The World Bank K. Subbarao is a Lead Economist (Social Protection) in the Human Development Department of the Africa Region. The views expressed herein are those of the author and do not necessarily reflect the opinions or policies of the World Bank or any of its affiliated organizations. Cover photo by Curt Carnemark, courtesy of World Bank Photo Library. Cover design by Tomoko Hirata. SYSTEMIC SHOCKS AND SOCIAL PROTECTION Contents Foreword v Executive Summary vii i Introduction 1 2 Rationale 3 3 Conceptual Issues 5 4 Design Features: Cross-Country Evidence 7 Design features that enhance program benefits to the poor 7 Cost-effectiveness 13 Implementation issues 15 5 Evaluation of Public Works Programs: Impacts on Poverty and Welfare 20 Targeting performance 20 Social gains 21 6 How to Plan, Implement and Evaluate: A Synthesis 22 List of Tables Table 1 Scale of Operations of Public Works Programs in Selected Countries 1 Table 2 Public Works: Program Wage (PW), Minimum Wage (MNW) And Market Wage (MW) in Selected Countries 7 Table 3 Proportion of Public Works Projects That Set Project Wages Below District-Level Market Wagesfor Unskilled Labour 9 Table 4 Cost-effectiveness of the Two Workfare Programs Under The Base-case Assumptions 15 Table 5 Marginal Odds of Participation for India's Main Antipoverty Program in Rural Areas 20 iV AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES List of Figures Figure 1 MEGS Labour Attendance 8 Figure 2 Seasonality of MEGS Employment 12 Figure 3 Designing and Implementing Public Works 23 SYSTEMIC SHOCKS AND SOCIAL PROTECTION V Foreword In much of Sub-Saharan Africa, poor households I welcome this study as an important overview are exposed to systemic risks ranging from nat- of the various issues involved in the design and ural disasters to sudden changes in economic implementation of public workfare programs. The policies. One immediate result of such systemic paper pulls together available evidence on design shocks is the loss of employment and livelihood. issues, cost-effectiveness, and evaluation of out- Depressed formal sectors in much of Africa offer comes and impacts of workfare programs. Based on little protection to workers in such times. AIDS is the evidence reviewed, the paper offers some good further exacerbating the condition of poor house- practice design features that can enhance the scope holds: as more breadwinners fall ill or die, the pres- for self-selection, enable a higher participation of sure on informal social protection networks women, lower transaction costs of participation by increases to the breaking point. In this context, the poor, and to a greater degree, involve the pri- public workfare programs assume great importance vate sector or non-governmental agencies in the as counter-cyclical interventions providing short- implementation of the program. It is hoped that the term employment to poor, unskilled workers who study will be of use not only to Bank staff engaged are impacted by adverse risks. in the implementation of multi-sectoral, program- Public workfare programs have been in operation matic lending operations but also to governments in many regions, including Sub-Saharan Africa, for interested in reforming their existing programs or several decades. Given its risk-prone environment, embarking on new workfare programs. Africa's extensive use of public workfare programs is not surprising. Governments in different contexts have resorted to such programs, both as self-stand- Birger Fredriksen ing programs as well as components of large multi- Sector Director sectoral interventions such as Social Investment Human Development Department Funds. Yet to date, there has not been a synthesis of Africa Region experience of public workfare programs in countries of Asia, Latin America and Africa. vi AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES Acknow1~~~~~~~~~~~~~~~4gments ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ W Fo mnyMpt ~n~n~ts izd ugetinsona erlerd~at,I oudlie o hnkJon lmquist, SYSTEMIC SHOCKS AND SOCIAL PROTECTION Vii Executive Summary P ublic works programs have been important ties. Fourth, the durable assets created will have the p counter-cyclical program interventions in potential to generate second-round employment developed as well as developing countries. benefits. Fifth, the program is highly amenable to In the developing world in general, and in Asia and geographic targeting. Finally, in many countries the Africa in particular, public works programs have program has helped the development of small-scale been significant policy instruments to mitigate the private contracting capacity. impacts of climatic risks on poor farmers and farm The success of the program depends very much laborers. These programs typically provide short- on the design features. A critical design feature is term, unskilled manual labor employment on proj- the level of the wage rate. Self-selection can be pro- ects such as road construction and maintenance, moted if the public works wage is slightly below the irrigation infrastructure, reforestation, and soil con- market wage for unskilled labor. Cross-country servation. They have been used to counter climatic experience reviewed in this paper suggests that risks in several countries, including Bangladesh, though there is much variation in the ability of gov- India, Ethiopia, Kenya, Zimbabwe, South Africa, ernments to fix a wage rate that is consistent with Tanzania, and Ghana. In Korea in 1997, public self-selection, several countries have adopted inno- works programs were the main program instru- vative approaches to facilitate wage setting that ment to counter financial risk-induced unemploy- promoted self-selection. The mode of wage pay- ment. Though known as "public" works programs, ment also influences the degree to which the pro- the actual implementation responsibilities are now gram is targeted to the poor in general and women being handled in several countries by small-scale in particular. In some African countries, women private contractors, NGOs, or social funds. favored task-based wage payment because it The rationale for public works programs rests on enabled them to dovetail household chores with six considerations. First, the program provides income-generating activities. income transfer benefits to poor households at crit- An important determinant of the cost-effective- ical times. Second, with good timing, the program ness of the program is the share of the wage bill in also confers consumption-smoothing benefits, thus total cost. Experience reviewed in this paper sug- countering the risk of consumption shortfalls gests that it is not easy to achieve high labor inten- during agricultural slack seasons/years. Third, sity even when known labor-based methods of well-designed workfare programs help construct production are available. Careful attention to detail the much-needed infrastructure and thus minimize is needed to attain a high labor intensity without the trade-off between public spending on income compromising the quality of assets created. Evi- transfer versus spending on developmental activi- dence bearing on cost-effectiveness of public works Viii AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES programs suggests that the program is highly rec- impacts of female empowerment have not been ommended when attention is paid to the quality of taken into account in the available estimates of ben- assets created, and when possibilities exist for such efits and costs. Also, when comparing the cost- assets to create second-round employment benefits. effectiveness calculations of workfare programs A word of caution is needed in interpreting the with other transfer programs, it is important to bear cost-effectiveness calculations cited in this paper. in mind the savings in administrative costs effected These calculations take into account only the trans- by self-selection, in addition to other factors. fer benefits conferred to the poor. The risk benefits The main constraint in implementing public of the program-the benefits of reduced risks due works programs in much of Africa is the lack of to consumption smoothing-are rarely factored capacity. This constraint can be eased if donor activ- into the calculations of cost-effectiveness. This is ities are coordinated and assistance is provided to one reason why the cost-effectiveness calculations build private contracting capacity. In all countries, often appear less favorable. Another limitation of and particularly in Sub-Saharan Africa, assured cost-effectiveness calculations is that only the direct funding, community participation, sound technical transfer benefits are factored into account. Indirect assistance, and proper understanding of societal benefits of workfare programs in terms of short- structures and communities where projects are and medium-term impacts on the rural market located can vastly improve the effectiveness of wage rate and the (indirect) social beneficial workfare programs. Systemic Shocks and Social Protection Role and Effectiveness of Public Works Programs Africa Region Human Development Working Paper Series 1 Introduction P ublic works programsl have been important the poor cope with co-variate risks associated with p counter-cyclical interventions in developed climatic and systemic shocks. Table 1 provides as well as developing countries over the last century In England, workhouse relief, to which the Table I able-bodied poor were restricted after England's Scale of operations of public works programs 1834 Poor Law Amendment Act, explicitly self-tar- in selected countries geted on the poor by aiming to provide pay and con- A. NATIONAL PROGRAMS ditions less eligible than the meanest available Person days of alternative (Himmelfarb 1984). Several Western Country employment created countries adopted different types of public works Botswana (1992-93) 7 million person days programs during the Depression years (1931-36) and again during milder recessions. In much of South Asia, public works programs began in the Kenya (1992-93) 0.6 million person days 1950s as "food for work" programs financed by food India National (1994) 800-900 million person days aid received from Western countries. These pro- grams now operate as govemment programs of cash Chile (1987) 4045 million person days for work, providing short-term employment at low Egypt (no date) 27-30 million person days wages. In Korea, the public workfare program was a Argentina (1998-2000) 400,000 persons core safety net during the financial crisis of 1997-98. Public works programs typically provide short- B. PROGRAMS IMPLEMENTED UNDER SOCAL INVESTMENT FUNDS term employment at low wages for unskilled and (MID-1980S TO EARLY 1990s) semi-skilled workers on labor-intensive projects Country and program Employment generated such as road construction and maintenance, irriga- Bolivia: FSE 731,000 person-months tion infrastructure, reforestation, and soil conserva- tion. Public works programs are now best viewed as Honduras: HSIF 140,000 person-months a means of providing income support to the poor at El Salvador: FIS 55,400 person-months critical times rather than as a way of getting the Peru: FONCODES 24,500 person-months unemployed back into the labor market. The les- sons from experience of public workfare are rele- Panama: FES 28,000 person-months and sons ~~~~~~~~~~~~~~~~~~~~~~~3,000 permanent jobs vant for all risk-prone countries generally, and for countries of Sub Saharan Africa and Asia particu- Nicaragua: FISE 73,000 person-months larly, because of their considerable potential to help Source: World Bank 1994: Subbarao 1997. 1 2 AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES information on the scale of the public works pro- little consultation with or involvement of communi- grams in selected developing countries. As can be ties or local governments in the selection and exe- seen from this table, in some countries the program cution of projects. Recent years have seen changes has been implemented on a national scale, whereas in the financing and implementation arrangements. in others the program has operated on a small scale In particular, in some countries, the scheme as one of the components of multi-sectoral inter- "provider" or "financier" (usually the government, vention, viz., Social Investment Fund. but also NGOs or international aid agencies such as There is much confusion about the meaning and the World Food Program (WFP)) has been sepa- scope of public works programs (also known as rated from the scheme "implementer," which could workfare programs) across countries. The term be the line ministries of the government, a private "public works" often creates an impression of a contractor, an NGO, or a Social Investment Fund. government-run program to create jobs. This was This paper is organized as follows. It begins with indeed the case in much of former Soviet Union and a brief discussion in Section 2 of the rationale of Eastern Europe, where public works programs were workfare programs in the developing world. Section understood to mean a "wage subsidy" program, 3 provides an overview of the conceptual issues. that is, the government paid the entire wage bill to Section 4 discusses the design features of a public encourage private entrepreneurs and state enter- workfare program, reviewing "real world" country prises to hire more workers and thus "solve" the experience with reference to each of the design fea- unemployment problem. This view is slowly chang- tures. Section 5 deals with implementation problems ing. Some Central Asian republics are trying to and financing issues and gives a brief overview of introduce public works at low wages as a short- available evaluations, with particular reference to term income transfer program for the poor along benefits and costs and to the distributional outcomes the lines of programs in South Asia. of workfare programs. The last section provides a In recent years, the implementation arrange- synthesis of how to plan, implement, and evaluate a ments for public works programs have changed, so public workfare program. the word "public" in "public works" has become somewhat inaccurate. "Old school" public works programs, typically financed and implemented by 1. In this paper, the terms "public works programs" and 'workfare programs" are used interchangeably. Both refer to the public works departments of central govern- programs in which participants must work to obtain benefits. ments, tended to display the drawbacks of other These programs offer temporary employment at a low wage centralized programs, namely, the creation of large rate, and have been widely used for fighting poverty. bureaucratic structures, lack of accountability, and 2 Rationale he argument for public works in nineteenth any policy intervention that lessens the risk of T century England centered around the ethic starvation for those surviving on the edge are to of work, often dirty and nasty work. The jus- be valued highly. tification for such public works during the Depres- sion years stemmed largely from the macro * Third, if well designed, the program can help imperative of restoring the aggregate demand. In construct much-needed infrastructure. For exam- many low-income countries today, the rationale is ple, the famous Maharashtra Employment Guar- vastly different from the motivations of programs antee Scheme of India, which has been in launched in the West in the past. operation for over three decades, has created In low-income countries, public workfare pro- considerable irrigation infrastructure and rural grams are undertaken with four objectives in mind: roads in the state of Maharashtra. Some of the durable assets created by the program have gen- • First, the program provides transfer benefits to erated (or can generate) additional second-round the poor. The transfer benefit is equal to the wage employment benefits. rate, net of any costs of participation incurred by the worker. In countries with high unemploy- * Fourth, the program is amenable to geographic ment rates, transfer benefits from a good work- targeting. Poor areas and communities can bene- fare program can prevent a worsening of fit directly from the program (in terms of transfer poverty, especially during periods of adjustment benefits), and indirectly from the physical assets or transition. created and/or maintained. To this extent, well- designed workfare programs can contribute * Second, depending on its timing, the program directly to enhancing the growth potential of less may also confer consumption-smoothing or sta- endowed regions. bilization benefits. The stabilization benefits arise mainly from the scheme's effect on the risk of a In addition there can be other (often unintended) decrease in consumption by the poor typically spin-offs from a public works program. For exam- during agricultural slack seasons. For example, if ple, workfare programs can build the capacity of the program's timing synchronizes with the agri- communities to manage their own affairs, strength- cultural slack seasons when the market demand ening local governments and other institutions. If for labor is low, workers are most likely to gain the design features are carefully thought through from the resulting income stabilization and con- (see Section Four), public works programs can sequent consumption smoothing. The benefits of encourage the participation and empowerment of 3 4 AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES women. Private contracting capacities could also run public works programs in many African coun- develop. In some countries, the private sector and tries, providing food as wages, the motivation being NGOs have been involved in the implementation of provision of household food security. public works programs. Finally, the links between Thus, in low-income countries, public works pro- workfare programs and household food security grams are undertaken with the multiple objectives deserves special mention. In much of Africa and of temporary income transfer benefits to the poor, South Asia, public works programs originated as consumption smoothing, household food security, "food-for-work" programs that paid wages in food, asset creation, and poor area development. The pro- usually donated food. Even now, some donor- grams are often regarded as vital for coping with driven programs (such as the World Food Program) climatic risks pervasive in much of Africa and Asia. 3 Conceptual Issues T he transfer benefit to a worker amounts to market wage for unskilled labor. A low wage is T the wage he/she gets from the scheme, net most likely to render the program unattractive to of the costs of participation (such as the cost the non-poor. A low wage will keep the overall par- of transport) and the earnings lost from alternative ticipation rate low and at the same time ensure that employment. If the costs of participation and a greater proportion of poor workers will participate income from alternative sources are negligible, and in the program than would be the case if the pro- if the program has no effect on the labor market and gram wage were higher. Given a budget, a low the structure of market wages, the transfer benefit wage would avoid job rationing. Thus, a low pro- approximately equals the program wage times the gram wage has several merits. However, a low duration of employment. In reality, these assump- wage rate will also result in lower transfer earnings tions are unlikely to hold. For example, the costs of per (poor) participant. participation and foregone earnings are only rarely Finally, in some situations the poor may incur zero. Most workers have to walk long distances to transaction costs that may further reduce the trans- the work sites or incur transport costs. In the fer benefit. For example, if the implementing agen- absence of the scheme, workers typically work for a cies and the institutional framework are subject to few days in alternative jobs, which they give up corruption and leakage, the poor who participate in when slightly longer-term employment is offered the program may have to a pay part of their wage to by a public works project. Moreover, unless the scheme organizers. This may further reduce the scale of the public works program is very limited, transfer benefit of the program. The particular the program is likely to put an upward pressure on implementation arrangements, the institutional the market wage rate. In that case, the net transfer framework, and the overall efficiency of scheme benefit (the direct program wage benefit plus the administration greatly influence the transfer bene- indirect benefit of an increment in market wage fits accruing to participants. resulting from the program) may be higher than the The longer the period of employment, the larger program wage. Thus, depending on the impact of the sum of total transfer earnings going to all partic- the scheme on the wage rate, the amount of the fore- ipants can be, and the higher the share of wages in gone earnings, and the costs of participation, the net total cost of the program will be. These two param- transfer benefit may be higher or lower than the eters-the share of wages and the duration of program wage. employment-vary depending on the nature of the In order to enable workers to self-select into the project. As mentioned in later sections, there is program, it is desirable to keep the program wage much cross-country variation in both the share of low, that is, somewhat lower than the prevailing wages and the duration of employment. 5 6 AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES The potential welfare gains from a public works holds often suffer precipitous shortfalls in con- program also depend on the source of financing. If a sumption and nutritional status in slack seasons public works program is entirely aid-financed, the and during periods/years of drought. A workfare transfer benefits to workers are a net addition to all program targeted in time and space to the regions other benefits flowing from programs funded out of most affected by monsoon failures or seasonal general tax revenues. However, if the program is to drops in economic activity can give poor partici- be funded out of general tax revenues, it would be pants "consumption smoothing" or "risk reduc- important to look at the counterfactual situation, tion" benefits. Income stabilization can prevent that is, what would have been the benefits to partic- acute distress and prevent poor households from ipants from alternative ways of spending the same onerous adjustments such as distress-selling of amount of budgetary resources. Also, it is useful to assets during years/seasons of crop failure. In other know if a public works program has been extended words, to poor households who lack avenues or at the expense of other activities that confer non-labor who cannot afford to insure themselves, the risk income to poor participants, such as education or benefits of a public works program can be as impor- hospital services. Rarely is it possible to evaluate tant as the transfer benefits. However, it is not this counterfactual element empirically, but in order always possible to implement the program when to understand the true benefits of public works pro- the poor are likely to sustain consumption short- grams, it is important to bear in mind the issues falls. For example, during periods of heavy rainfall arising from the source of the program's financing. when all economic activities come to a halt, logistics The stabilization benefit of the program reflects may not permit implementation of a public works the program's "insurance" function. It depends on program, especially in hinterland villages. There- the timing of the program. In predominantly agri- fore, a workfare program may not be the appropri- cultural societies, household incomes move up or ate instrument to protect the poor during the down depending on seasonal activities. Poor house- "hungry" season. 4 Design Features: Cross-Country Evidence T his section discusses the design features of a Table 2 public works program using three broad cri- Public works: Program wage (PW), teria: (a) how to self-target the program to minimum wage (MNW) and market wage (MWt the poor and maximize the benefits to participants, (b) how to improve the cost-effectiveness of the pro- PW in relation to gram, and (c) how to implement the program. The Country/Program MNW and/or MW first two sub-sections cover only the major national- Bangladesh: level programs, whereas the third sub-section Cash For Work, 1991-92 PWMW 1 ~~~(b1) MEGS: up to 1988 PW=MNWMW Pakistan: Design features that enhance program benefits IGPRAa isa, 1992 PWMW The level of the wage rate Food For Work 1987 PW*>MW Botswana: The wage rate is a key element in determining Cash For Work PWMW the degree to which the poor self-select into the pro- Kenya: gram thus determining the distributional outcomes Cash For Work 1992-93 PW=MNW>MW of the program. In order to promote self-selection, it Chile: is best for a public works program to offer a wage Cash For Work 1987 PW 1990-91 engineers were averse to adopting labor-based 12 b methods, largely because they were unfamiliar with labor-based methods of production and because ~ 10 - - ____ ~ extra supervisory inputs (expenses) are needed as § 8 _ the size of labor gangs increase (Adato and others . -- /f 1999). Where the work has been entrusted to private t 6 -- / contractors, the outcome with respect to labor 4 - :. . -~ intensity is unpredictable. Evidence from Ghana 2 suggests that timely availability of project funds is 0- important for the adoption of labor-based methods. Ns^ _10- # 0 .c Where a program was financed with donor funds and payments were quick and guaranteed, contrac- SYSTEMIC SHOCKS AND SOCIAL PROTECTION 13 tors would resort to labor-based methods. Where an additional determinant of the cost-effectiveness the program was funded by government funds, of the program. Ravallion (1999:34-35) defined which were uncertain and often arrived after enor- these five variables as follows: mous delays, contractors did not favor labor-based methods, fearing strikes for delayed wage pay- (a) Budget leverage. The government can ments (Subbarao and others 1997). The incentives require co-financing from nonpoor neigh- worked exactly in the opposite way in Argentina. If borhoods for subprojects that will benefit private contractors were in charge and had to meet them. Let government (central plus local) contract standards, they were unlikely to choose spending be G, and let this spending be labor-intensive methods. In Argentina, the federal leveraged up to result in a total budget of government paid only the wage cost. The payments G + C, including private co-financing (C). to workers were generally on time and not subject to any significant problem. The issue was more the (b) Labor intensity. Some of the participants availability of materials (non-wage inputs). So may not be poor, so let the share of all municipalities, particularly the poor ones, had an wages paid in total operating cost be incentive to choose more labor-intensive projects (W + L)/(G + C), where W is the wage because of the difficulty in obtaining non-wage received by the poor and L denotes leakage inputs. The challenge for project management was to the nonpoor. to get the labor to do projects that had a reasonable value for communities. (c) Targeted labor earnings. This is the propor- Development and dissemination of labor-inten- tion of the wages paid out to poor workers, sive designs, coupled with quick payments, can W/(W + L). encourage implementing agencies to adopt labor- based methods. Where works are entrusted to (d) Net wage gain. This is the share of the gross private contractors, innovative incentive systems wage received by the poor after subtracting need to be developed to promote labor-intensive all cost of participation, including income methods. forgone from other work. The net wage The reviewed experience suggests that some gain is NW/IW, where NW stands for wages countries managed to incorporate many of the ideal net of forgone income or other costs of par- design features of a public workfare program (see ticipation. Box 1). Much can be accomplished if countries are aware of these ideal design features before launch- (e) Indirect benefit. Let IB denote the indirect ing a workfare program. benefits to the poor, such as when the assets created are local public goods in poor Cost effectiveness neighborhoods. Four variables determine the cost-effectiveness Ravallion's simulations (Table 4) show the cost of of public works programs. These are labor intensity transferring $1 of income to the poor for a typical (the proportion of total wage bill going to poor middle-income country (with a poverty rate of 20 workers), targeting performance, net wage gain percent) and for a typical low-income country (with (gross wages minus all costs of participation a poverty rate of 50 percent). If only current benefits incurred by workers), and the indirect benefits are considered, the cost to transfer $1 of income to flowing from the assets created. In some countries, the poor is $5.00 for middle-income countries and governments require co-financing from nonpoor $3.60 for low-income countries. But if future gains communities for implementing sub-projects that from the assets created are also included in the ben- benefit those neighborhoods. In such instances, the efits, the cost of transferring a $1 gain to the poor budget leverage, that is, the share of the govern- drops to $2.50 for both middle- and low-income ment's outlay that actually benefits the poor, can be countries. Nevertheless, at first sight, it appears that 14 AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES Box I What are the key designbfA4"04es oagodpbIt wors romr To realiz6eJte ull poeta f a workfere porMMa a8 povrtPeM cigan rs oopng intuet t Ircmmmended that: * The wage rate should be set no higher than the prpevaiig~ Mare waePruIkdmna labo in ft fhe etn int which the schemne is * Restrictios on elgbilbity should be avoided; the fact that an indiida wat oka hswg aeshouldidel be the onl reqwre * IfratiningIs equied (ecaue dmandfor orkexceds te bugetavaiabl~at te wge st), he rogrmtshuld e trgtto popoo areas, as indcatedO bya rd"ibl vertmp M oweer fexbiit shud ealoe 9otr bugt aloaln ars ra to reflct differences in demiand for fthschee * The labor Intesit (shOa of wage biR inW toa cotsould beat iha*osba~Telvlo ao intensWity w 00idend on the rltive importnceatachd6tw mmdat1ncm gains versus1(incme n6 te)gist h orfo the asst created. Thi wilvary from setting to seting * The projets shoud betargete to poor~ aesadsoltrtonuethtt sescetdaeof maxmu valu to poor popdle In thos aras.My ecepion-in hic th assts argly bneft te no-por-souldreqireco-inancin from tebeneficiaries, and this oe s0hohUldgo back intothe budgtof tescee *In ordert Co ecurage*fmale~ pridipation, th aprpit omo ae siprat Or exape vwomna benefit frmpjieertes or taskbased wages Somtie wagesin hefrofod av atra ted mWeom to PA sites Al,provso of chIlda or pre- school serviWccan iprovepartcpto by wmen * Transacdtio costs to th poor shouldi be kept lowQ One imprtntow mastacopihhsi blOcaing. projec tsite close to villages. It is Als necessary to ensure appropriate mdiation of. NGs for prtcigThrightsd of te. poor in reation to pgrogrm managers * The program should include anastmaintenande component. a public works program is not an inexpensive way program more cost-efficient than other targeted to transfer income and consumption-smoothing programs. Second, other targeted programs imple- benefits to the poor. mented via administrative targeting may also have it is important to bear in mind some of the limi- high costs of leakage. Subbarao and others (1997) tations of the cost-effectiveness calculations and estimate the leakage to the nonpoor from targeted associated simulations. First, it may be helpful to food programs in several developing countries and generate similar numbers for other programs and suggest that the proportion of total transfer benefits compare the cost-effectiveness ratios across pro- to the poor ranges from 19 to 93 percent across grams and countries. When such a comparison is countries. For food subsidy programs in India, Rad- done, important limitations of the cost-effectiveness hakrishna and Subbarao (1997) estimate the share of calculations of workfare programs in Table 4 will expenditure reaching the poor to be 16 to 19 per- become evident. First, to the extent that a well- cent. For housing subsidies in various countries, the designed public works program is self-targeted and same source estimated the share going to house- does not incur administrative costs of targeting, the holds below the median income to be between 10 cost savings in public works may render the and 50 percent, implying very high levels of leakage SYSTEMIC SHOCKS AND SOCIAL PROTECTION 15 Table 4 works programs offer a reasonable degree of guaran- Cost-effectiveness of the two workfare tee of employment during off-peak seasons, it is pos- programs under the base-case assumptions sible that market wage rates for unskilled labor may Middle-income Low-income increase as a result of the higher reservation wage Country (poverty Country (poverty induced by the public works program. There is only rate=20%) rate=50%) one study that estimates these indirect gains for the Budget leverage: (G+C)IG 1.0 1.0 Maharashtra Employment Guarantee Scheme Labor intensity: (W+L)/(G+C) 0.33 0.5 (MEGS) (Gaiha 2000). With an analytical model, the Targeting: W/(W+L) 1.0 0.75 study tracks the interdependence between the agri- Net wage gain: NWNW 0.6 0.75 Poor people's share of cultural market wage, the MEGS wage, and non-farm total benefits: IB/SB 0.2 0.25 wages. The study finds that the program has had sub- Benefit/cost ratio: SB/(G+C) 1.0 0.5 stantive indirect benefits: if MEGS wages rose by Rs.1, Current + future gains to the poor rural farm wages would increase by Rs.0.17 in the per $ of spending: B/G 0.40 0.41 short run and by Rs.0.28 in the long run. Cost of $1 gain to the poor $2.50 $2.50 Finally, the future benefits of a public works pro- Current earnings gain per $ of gram could be substantial. A recent study for MEGS program spending: CB/G 0.20 0.28 noted that the benefits in terms of asset creation to Cost of $1 extra current eamings $5.00 $3.60 the rural economy of Maharashtra have been con- siderable (Gaiha 2000). Source: Ravallion (1999). To sum up, cost-effectiveness calculations are important, but the numbers need to be interpreted with caution. On the cost side, it is important to of benefits to the non-poor. Thus, available esti- bear in mind the implicit savings in costs induced mates of leakage and administrative costs suggest by self-selection. Both the direct costs and the that after a careful comparison, one might well find opportunity costs need to be factored into the cal- that other programs might do a lot worse than public culations. On the benefits side, the failure to recog- works programs in terms of cost-effectiveness (see also nize the indirect and second-round benefits as well section 5, table 4). as the direct ones may underestimate the benefits, A third limitation of cost-effectiveness calcula- yielding unfavorable cost-benefit ratios. Thus, a tions as shown in Table 4 is that the benefits esti- simple cost-effectiveness calculation of a works pro- mates consider only the transfer benefits. The risk gram that does not take into account the above fac- benefits, that is, the benefits of reduced risks due to tors might show the program to be expensive, consumption smoothing, are rarely factored into compared to other transfer programs. But after calculations of cost-effectiveness. We have noted taking into account both the direct and indirect that these risk benefits may be extremely important benefits, the program may well tum out to be cost- for poor people who lack access to risk-coping effective. instruments or who cannot afford to insure them- selves against potential risks of income or con- Implementation issues sumption shortfalls. If work is easily obtained at sites closer to the homes of participants, workfare Although labor-intensive public works programs programs can be more responsive than most other have the potential to create short-term employment safety net programs to risks of sudden shortfalls in to improve risk management by poor households, consumption of poor households. several implementation issues may arise. Institu- Fourth, the cost-effectiveness calculations shown tional capacity for designing and implementing in Table 4 take into account only the direct transfer public works programs is quite diverse across coun- benefits. Indirect benefits in terms of short- and tries. Countries such as India and Bangladesh have medium-term impacts of the program on the rural over time built within-country capacity to imple- market wage rate are rarely evaluated. If public ment public works programs, especially during 16 AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES periods of crop failures. Within-country capacity to tor has overall responsibility for coordinating the implement workfare programs is somewhat limited work and allocating funds among blocks within a in African countries. International agencies (WFP, district. ILO, bilateral agencies) have been active in public Participatory assessments of both programs in works programs in many African countries. India's largest state of Uttar Pradesh have pointed "Although recorded on efficiency and effectiveness to poor implementation of the program. Decisions appears to be mixed, some successful examples are on who participates in the program are made by the emerging" (von Braun, Teklu and Web 1992:31). The elected officials at the village level, who often major implementation issues are how funds flow, excluded members of some of the socially deprived how projects are selected, how much self-selection communities in some provinces. The poor partici- participants practice, and how well the program is pant's perceptions of the main objective of public monitored. works programs differed from the professed objec- Among the world's largest and best known rural tives of these programs. For example, most poor and urban public works programs are India's participants perceived the JRY and EAS as infra- nationwide Jawahar Rojgar Yojana (RY) and the structure projects to build roads and bridges and Employment Assurance Scheme (EAS), both not as programs intended to provide employment designed to help ensure gainful employment for and income support to the poor to prevent off-sea- poor households and to contribute to creation of sonal shortfalls in consumption. These findings rural and urban infrastructure. The two programs from the participatory evaluation suggest that, are administered very differently. The JRY is admin- regardless of the differences in the flow of funds istratively complex. Of the total fund made avail- and the differing approaches of the two programs able, 75 percent is earmarked for various rural (supply-driven versus demand-driven), implemen- infrastructure schemes; and the rest goes to social tation problems persist even in the country with the forestry projects. All, however, are allocated to the longest experience. states according to a set formula based on the pro- The implementation experience of the Trajabar portion of poor persons residing in each one. States workfare program in Argentina has been positive. in turn allocate JRY funds to districts based on pop- Like the Indian programs, it is a large-scale under- ulation shares and an index of backwardness. From taking, but several factors have enabled a more there JRY funds pass down to block and village effective implementation. In Argentina, there were level strictly according to population shares. very clear and transparent guidelines from the cen- Unique in its decentralization-work plans and con- tral government, leaving implementation details in a tracts are administered by the village panchayats decentralized fashion to local and municipal (elected local bodies), subject to the basic program authorities. The project staff were highly commit- guidelines and overall clearance of the work pro- ted, funds were distributed across municipalities gram by the District Rural Development Authority following transparent and objective criteria (accord- (DRDA)-the JRY generates an estimated one bil- ing to the distribution of the poor unemployed), lion person days of employment each year, an and there was a sharp focus on monitoring and achievement which translates into the likely partic- evaluation so that problems could be identified ipation of some 30 to 40 percent of potential benefi- early and dealt with. Proven project evaluation and ciaries (World Bank 1998:38). Text should resume supervision procedures were adapted from those of flush left. Thus the program is largely supply- social funds and were successfully implemented. To driven, implemented by local bodies but subject to further enhance transparency, the selected projects pre-determined centrally-driven guidelines. In con- were published, as were lists of selected project trast, the EAS is demand-driven. District authorities beneficiaries. apply directly to the central administration for In some Latin and Central American countries, funding, and allocations are made based on the size including Bolivia, Honduras, El Salvador, Peru, and backwardness of blocks included in each dis- Panama, and Nicaragua, small-scale public works trict. In the EAS, unlike the JRY, the District Collec- projects have been implemented under the aegis of SYSTEMIC SHOCKS AND SOCIAL PROTECTION 17 Social Investment Funds (SIFs). Unlike India's JRY In Africa too, quite a few public works projects and EAS programs, these projects are submitted by have been financed and sponsored by Social Invest- the communities and thus are demand-driven. The ment Funds. By 1998, there were 19 public works SIFs screen the project proposals to determine via- projects implemented through social funds in 12 bility after making a quick cost-benefit analysis. The African countries, including Senegal, Guinea- communities themselves implement the approved Bissau, Madagascar, Mauritania, and Ghana projects, albeit by hiring private contractors. Many (Frigenti and Harth 1998). Over half of these have focused on social infrastructure such as school projects were supported by several donors in addi- and hospital building and repair. Most created tem- tion to the World Bank. The principles are the same porary jobs lasting five to six months. In Bolivia, as in Latin America, with project proposals pre- under the Emergency Social Fund (ESF)-the oldest pared by communities and submitted to the SIF for of the SIF interventions-the implementation and approval and financing. In Mauritania, for example, targeting performance have been good: 77 percent of neighborhood and town meetings determine prior- the participants came from the poorest 40 percent of ities and select projects. Because the programs seek the population. After the intervention, the average to work with small-scale private construction con- worker improved his income by 67 percent (orgen- tractors, the autonomous status of implementing son, Grosh, and Schachter 1992). However, regional agencies has proved to be important for processing targeting in the initial stages was not pro-poor, bids rapidly, awarding contracts quickly and with- largely because the poorest regions lacked the capac- out political interference, and accelerating pay- ity to demand projects. But this deficiency was cor- ments to contractors. In general, SIF-led project rected as the program improved its geographic preparation was very transparent and free from targeting. The ESF won praise for speedy and effi- cumbersome government procedures, although cient implementation and for significant benefits to there are some cross-country differences. In Sene- the economy. Each ESF job created an additional 1.1 gal, AGETIP's success in implementing public jobs in the economy, and the projected rates of return works was largely due to its insulation from clien- on approved projects was 22 percent. Following the telistic politics (Marc and others 1995). In contrast, successful launch of ESF in Bolivia, a number of Ghana's public works project was integrated into other Latin American countries (Honduras, El Sal- government ministries and consequently faced vador, Peru, Panama, and Nicaragua) introduced delays in implementation. SIFs. Financing infrastructure through small-scale It is worth stressing that public works activities public works has been one of the main components under SIFs have always been typically small-scale; of SIFs in all countries. Although employment gen- hence the implementation experience of these proj- eration was not the main motivation behind SIFs, ects is not strictly comparable to national programs many infrastructure projects did generate tempo- in Chile, India, Korea, Indonesia, or Argentina. rary employment lasting five to six months in some Line departments of governments in some of the poorest communities. Nonetheless, all SIFs African countries have implemented public works had to confront one major implementation prob- projects on a large-scale, albeit with donor funds. lem-they were not responsible for the execution of The main motivation of these projects is to provide the projects. The choice and implementation were food security during periods of crop failure (in left entirely to the communities. Because very poor Ethiopia, for example). In implementing such proj- communities lacked the capacity to develop proj- ects, however, evidence suggests that the public ects, in most SIF programs, they received fewer proj- works departments of governments typically ect benefits than the relatively better endowed favored equipment-based methods because they communities. However, the SIFs did provide an were perceived to be superior to labor-intensive opportunity for communities to register their methods and completed the work faster. It is quite demands and influence project selection. Moreover, possible that in some countries, equipment-inten- the more recent SIFs have begun to address the issue sive projects may offer greater opportunities for of building capacity in poor communities. rent seeking (Stock 1996). 18 AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES In some countries, such as Ghana, the task of ference is that the implementation in Africa rested implementing public works was entrusted to pri- largely with government departments (the Ministry vate contractors. Evidence suggests that contractors of Public Works in Kenya, the Ministry of Local have also been reluctant to adopt labor-intensive Government and Lands in Botswana, and the Prime public works, largely because of the complexities of Minister's office in Tanzania). Because programs managing large labor gangs (Stock 1996). In addi- were implemented largely by government agencies tion, in projects financed by the government, delays as part of their routine work program, little atten- in payments often discouraged private contractors tion was paid to such details as timing of the pro- from adopting labor-intensive methods. Contrac- gram, monitoring, quality of roads built, and so on. tors feared workers would strike if their wages For example, in Tanzania, the timing of works failed were delayed due to late arrival of government to synchronize with agricultural slack seasons. In all funds. Labor-intensive methods were more favored programs, wages are set uniformly regardless of in donor-funded projects where contractors re- type of work, location of work site, and variations ceived prompt payments. This contrasts with the in the work force (age, sex, education, experience, experience of Argentina, where no such constraints and so forth) (Teklu 1994). "The restrictive structure were allowed to emerge. of pan-territorial wage rates limited the flexibility of In Zimbabwe, two major programs operated. the road programs to adjust wage rates in accor- The Food-for-Works program replaced free distri- dance with local labor supply conditions" (Teklu bution of food in 1989. Villagers themselves selected 1994:22). Teklu (1994) also notes that the long-term the projects. Local food security was the primary benefits of public works projects are much dimin- goal, so that other aspects of the program such as ished in Tanzania due to poor maintenance of the quality and maintenance suffered. A second pro- assets. gram of public works began a few years later with One question that often arises in the implemen- the primary objective of development. Technically tation of a workfare program is who should be con- qualified persons have begun the program with sidered eligible to participate in the program. Some materials and tools provided by the government. In countries have laid down specific criteria. In Korea, Zimbabwe, there has been much greater commu- for example, only one member per household, the nity participation than is often the case, and the head of the household (usually male) can partici- demand-driven nature of projects should be com- pate. When the household head is receiving an mended. unemployment benefit, the spouse (female member Experience gained thus far suggests that financ- of the household) is not allowed to participate. In ing arrangements and the flow of funds influence other words, although the low wage rate could have critical design features, including the labor intensity promoted self-selection, the system was not of programs. The financing arrangements in African allowed to operate because of additional restric- countries are different from those found in South tions on participation. Once the wage rate fixed is Asia and Latin America in three key ways. First, in low enough to promote self-selection, the decision much of Africa, public projects are donor funded; of whether or not to participate, and who should domestic country contributions are negligible. In participate (male or female member of the house- Tanzania, for example, the UNDP and ILO jointly hold or both) should be left to the household. Often supported a labor-intensive public works program many countries do not follow this principle. (road construction and maintenance), with contri- Another problem is lack of capacity in many butions from The Netherlands, Germany, and Den- African countries. When programs are donor- mark. The government of Tanzania contributed 11 funded and short-term (typically for three or four percent of the total cost. The early programs in months following a drought), domestic capacity is Kenya and Botswana were also heavily donor- unlikely to be built. In this respect, the experiences funded. The second major difference is that, in of India and Bangladesh are worth contrasting with Africa, the provision of capital budgets by donors the experience of African countries. In both was often tied to technical assistance. The third dif- Bangladesh and India, most public works projects SYSTEMIC SHOCKS AND SOCIAL PROTECTION 19 operate throughout the year, albeit with seasonal ation, expansion, contraction, or dissolution. Far ups and downs in coverage. As a result, much from being a guarantee, or being there when domestic capacity is created over time. When capac- needed, public works have operated in much of ity is unknown, which is often the case in some Africa only when donor funding was available. countries, designers of public works programs may Fragmented coverage and weak capacity to find it useful to initiate a pilot phase during which respond in times of need undermine the credibility the implementing agency's capacity can be tested of public works programs to perform an insurance and lessons drawn. Projects that rely chiefly on gov- function for the poor of Africa. In all countries, and ernment agencies and ministries for sub-project particularly in Sub-Saharan countries, assured identification and implementation are most likely to funding, community participation, sound technical suffer from delays and limited capacities (technical, assistance, and proper understanding of societal administrative, financial, and participatory). structures and communities where projects are When many donors implement different public located can vastly improve the effectiveness of the works programs in the same country, lack of coor- workfare program as a risk-mitigating intervention. dination unduly stretches scarce administrative capacity, so that the coverage of these programs is often neither extensive nor deep. Most public works 1. The main reason is that the detailed design features and cost-effectiveness aspects of workfare programs implemented projects in Africa do not have clear criteria for initi- under the rubric of Social Investment Funds are not available. 5 Evaluation of Public Works Programs: Impacts on Poverty and Welfare E valuations of workfare programs in general, useful safety nets during a macro economic crisis. E and of MEGS in particular, suggest that it is Survey-based impact evaluation methods were possible for governments to use workfare to used to assess the gains to participating workers bring significant transfer and stabilization benefits and their families from the Trabajar program in to poor households. Argentina. Propensity-score matching methods were used to construct a comparison group to Targeting performance survey Trabajar participants from an identical national sample survey implemented at the same As for targeting outcomes, nearly 100 percent of participants in Chile's public works program belonged to poor households. In India (in both the Table 5 nationwide program and MEGS) and in Argentina's Marginal odds of participation for India's Trabajar program, 60 to 70 percent of households main antipoverty program in rural areas participating were poor. India's National Sample Public Integrated Rural Public Survey results for 1993-94 contain data about works Development Distribution household participation in three key safety net pro- Quintile programs Program System grams: public works schemes the Integrated Rural Development Program (IRDP, a micro finance pro- (poorest) (3.27) (15.49) (8.14) gram), and the Public Distribution System (PDS), a food subsidy program. By collating the data on pro- 2 0.93 1.28 0.99 gram participation with the data on total consump- (3.64) (17.73) (7.26) tion expenditure per person at the household level, 3 0.80 1.21 0.91 it is possible to determine the current distribution of (2.98) (23.52) (6.88) benefits from public spending across income 4 0.92 0.96 0.86 groups for the programs. Research by Lanjouw and (4.32) (19.09) (7.16) Ravallion (1998) shows that the poorest quintile is 5 0.55 0.39 0.81 well served by public works programs, with the (3.29) (8.06) (6.27) credit program (IRDP) in second place, and the food Note: The table gives the instrumental estimates of the regression coefficients of the subsidy program (PDS) doing least well in reaching quintile-specific program participation rates across regions on the average rate by state the poor (Table 5). for that program. The leave-out mean participation rate is the instrument for the actual Even in middle-income countries such as ~~~mean. The numbers in parentheses are t-ratios. Source: Laniouw & Ravallion (1999:269). Calculations based on the 1993-94 National Argentina, workfare programs have served as Sample Survey. 20 SYSTEMIC SHOCKS AND SOCIAL PROTECTION 21 time. Income gains were then estimated by compar- ultrapoor families (Subbarao, Ahmed, and Teklu ing incomes of the Trabajar participants with those 1996). of the matched comparison group. The results have indicated that Trabajar jobs were well targeted to Social gains the poor. For example, 80 percent of participating workers come from families with an income per It appears that well-designed workfare programs capita that places them among the poorest 20 per- have the potential to confer significant social gains. cent of the national population and 60 percent come For example, India's Maharashtra Employment from the poorest decile (Jalan and Ravallion 1999). Guarantee Scheme (MEGS) was designed to A good evaluation of the experience of public encourage the participation of women. Employ- works in South Africa is now available (Adato and ment was provided within five kilometers of the others 1999). The outcomes on targeting perform- place of residence, creche facilities were provided, ance appear to be somewhat mixed. Using socio- and male-female wage discrimination was elimi- economic data at the district level, the study found nated. Close to half of participants were women. that some districts with very high levels of poverty Datt and Ravallion (1992) have quantified the and unemployment had no projects, while some impact of the program and found that the severity with low levels of poverty had received several. of poverty has fallen from 5.0 percent to 3.2 percent, Though women were among the main target owing to participation in the MEGS. In addition to groups, only 23 percent of employment generated economic (transfer) gains, Dev notes social gains: went to women. "The MEGS also discourages sexual barriers and In The Philippines, largely because of a relatively inequality.... Women now dress better and their high wage rate (cash plus in kind), participants in economic power has given them a better status in public works projects appeared to come from mar- their families" (Dev 1996:54). ginal poor and non-poor families rather than from 6 How to Plan, Implement, and Evaluate Public Works Programs: A Synthesis F igure 3 provides a schematic view of the con- fits from the program and its targeting siderations that should be borne in mind at effectiveness. different stages in the planning of public workfare programs. (b) The timing and duration of employment First, it is important to begin with adequate often determine the stabilization gains from sources of financing. If funding is from general tax the program. It is important to remember revenues, it is useful to consider competing that even if transfer benefits are small, the demands for the generation of public goods vital for program's stabilization (risk) benefits could the welfare of poor households. Second, technical be large, especially in economies subject to feasibility, wage rate and mode of payment, choice periodic natural disasters. of projects, and community involvement must be considered. Third, it is necessary to consider how (c) The program can be designed to attract the choice of projects and the wage rate will affect more women participants, achieve a greater targeting effectiveness in particular and the labor degree of involvement by the private sector market in general. Fourth, the program's potential or non-governmental agencies, and lower for stabilization gains and improved risk manage- the transaction costs of participation by the ment of poor households needs to be borne in mind. poor. Finally, it is important to consider the distributional impacts, second round employment effects, gender (d) In order to achieve an acceptable level of impacts, and cost effectiveness. Other spin-offs such cost-effectiveness, it is very important to as community mobilization, women's empower- pay attention not only to the level of the ment, and other social gains need to be considered. wage rate but also to the degree of labor In designing and implementing the program, intensity, the quality of the assets created four general conclusions from the reviewed experi- and extent to which the poor actually bene- ences need to be borne in mind: fit from the created assets, and the second- round employment benefits arising out of (a) The level of the wage rate is critical for the assets created. determining both the distribution of bene- 22 SYSTEMIC SHOCKS AND SOCIAL PROTECTION 23 Figure 3 Designing and implementing public works Stages What to consider Needed: Source of financing: If out of general tax revenues, consider competing Capital demands from other public goods Food/Cash Consider other demands on scarce institutional and administrative Implementing Agencies recources Institutions Choice of projects, community involvement Public Works Projects Technical feasibility, labor intensity Level of the wage rate, mode of payment Targeting effectiveness t Labor market effects Immediate Impacts: Transfer gains Employment T Stabilization gains, improved risk management Cost-effectiveness ' r Distributional impacts of assets created Medium-Term Impacts: Assets Second round employment effects Qualitv of assets, asset maintenance Gender impacts, women's empowerment Other Spin-Offs P. o. Food security, improved nutrition Community mobilization References Adato M., L. Haddad, D. Horner, N. Ravjee, and R. Hay- Jorgensen, Steen, Margaret Grosh and Mark Schacter wood. 1999. "From Works to Public Works: The Per- 1992. "Easing the Poor Through Economic Adjust- formance of Labor Intensive Pubic Works in Western ment: The Story of Bolivia's Emergency Social Fund". Cape Province, South Africa." International Food Regional Studies Program, World Bank, Washington, Policy Research Institute. Mimeo. D.C. Datt, G., and M. Ravallion.1992 "Behavioral Responses to Lanjouw, Peter, and M. Ravallion. "Benefit Incidence, Workfare Programs: Evidence for Rural India." LSMS Public Spending Reforms, and the Timing of Program Working Paper, World Bank, Washington. D.C. Capture" The World Bank Economic Review: May 1999, - "Transfer Benefits from Public Works Employ- pp.257-273. ment: Evidence for Rural India." Economic Journal 104 Lipton, M. "Successes in Anti-Poverty." 1996. International (1994): 1346-69. Labor Office, Geneva, Discussion Paper 8. Mimeo. Dev, S.M. India's MEGS: Lessons from Long Experience. Radhakrishna, R., and K. Subbarao. 1997. India's Public 1994. Indira Gandhi Institute of Development Distribution System: A National and International Per- Research, Bombay. Mimeo. spective. World Bank Discussion Paper 380. Washing- Frigenti, L., and A. Harth.1998. Local Solutions to Regional ton, D.C. Problems: The Growth of Social Funds and Public Works Ravallion, M. "Appraising Workfare" The World Bank and Employment Projects in Sub Saharan Africa. World Research Observer, 14(1) February 1999, pp. 31-48. Bank, Africa Region. Washington, D.C. 2000. "Monitoring Targeting Performance when Gaiha, R. 2000. Rural Public Works and the Poor: A Review of Decentralized Allocations to the Poor Are Unob- the Employment Guarantee Scheme in Maharashtra. Fac- served". The World Bank Economic Review, May 2000, ulty of Management Studies, University of Delhi. pp 331-345. Mimeo. Stock, E., and J. de Veen. 1996. Expanding Labor-Based Himmelfarb, G. The Idea of Poverty. 1984. London: Faber. J. Methods in Road Programs. World Bank, Africa Techni- Jalan, and M. Ravallion. 1999."Income Gains to the cal Department, Washington, D.C.. Mimeo. 1996. Poor from Workfare: Estimates for Argentina's Traba- Subbarao, K. 1993. "Interventions to Fill Nutrition Gaps jar Program" Policy Research Working Paper 2149, at the Household-Level: A Review of India's Experi- Development Research Group, World Bank, Washing- ence", in B. Harriss, S. Guhan, and R.H. Cassen (eds.) ton DC. Poverty in India: Research and Policy. Bombay: Oxford Lanjouw, P., and M. Ravallion. 1998. Benefit Incidence and University Press, pp.231-258. the Timing of Program Capture. Development Research Subbarao, K. 1999. "Financial Crises and Safety Nets: Old Group WVorld Bank. and the New Poor in Korea", World Bank. Washington, Jalan, Jyotsna and M. Ravallion 1999. "Income Gains to D.C. Mimeo. the Poor from Workfare: Estimates for Argentina's Tra- bajar Program". Policy Research Working Paper 2149, Development Research Group, the World Bank, Wash- ington, D.C. 24 SYSTEMIC SHOCKS AND SOCIAL PROTECTION 25 Subbarao, K., A. Ahmed, and T. Teklu. 1995. Philippines: Von Braun, Joachim, T. Teklu, and P. Web, 1992: "Labor- Social Safety Net Programs: Targeting, Cost-Effectiveness Intensive Public Works for Food Security in Africa: and Options for Reform. Discussion Paper No. 317, Past Experience and Future Potential." International World Bank Discussion Paper No. 317. Washington, Labor Review, 13(1): 19-33. D.C. Walker, T., and J. Ryan. 1990. Village and Household Subbarao, K., A. Bonnerjee, J. Braithwaite, S. Carvalho, K. Economies in India's Semi-Arid Tropics, Baltimore, Md. Ezemenari, C. Graham, and A. Thompson 1997. Safety The Johns Hopkins University Press. Net Programs and Poverty Reduction: Lessonsfrom Cross- World Bank, Reducing Poverty in India: Options for More Countnj Experience. World Bank, Directions in Devel- Executive Public Services. Washington, DC, 1998. opment, Washington, D.C. World Bank, "Poverty Alleviation and Social Investment Subbarao, K. "Public Works as an Anti-Poverty Program: Funds: The Latin American Experience", The World An Overview of Cross-Country Experience", Armerican Bank, Latin America Human Development Depart- Journal of Agricultural Economics, 79 (May 1997): 678- ment, Washington, DC, 1994. 683. Teklu, T. 1994. Labor-Intensive Rutral Roads in Kenya, Tanza- nia and Botswana: Some Evidence on Design and Practice, International Food Policy Research Institute, Washing- ton, D.C. THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. Telephone: 202 477 1234 Facsimile: 202 477 6391 Internet: www.worldbank.org E-mail: afrhdseries@worldbank.org In much of Sub-Saharan Africa, poor households are exposed to risks ranging from natural disasters to sudden changes in economic policies. One immediate result of such shocks is the loss of employment and livelihood. In this context, public workfare programs, which typically provide short-term, unskilled manual labor employment on projects such as road construction and maintenance, have been important counter-cyclical interventions in many regions, including Sub-Saharan Africa for several decades. In the developing world in general, and in Africa and Asia in particular, public works programs have been significant policy instruments to mitigate the impacts of climatic and systemic risks on poor farmers and unskilled and semi-skilled workers. Systemic Shocks and Social Protection: Role and Effectiveness of Public Works Programs discusses the rationale of workfare programs in the context of social risk management, outlines the design features of a "good practice" workfare program, and provides an overview of African and Asian experiences with respect to a number of design features, including the wage rate, labor intensity and project selection and implementation. The paper also reviews estimates of cost-effectiveness and available evaluations of effectiveness, and finally draws some lessons from experience. It is hoped that the study will be of use not only to Bank staff but also to governments interested in reforming their existing programs or embarking on new workfare programs.