37792 267 Human Development October 2006 Findings reports on ongoing operational, economic, and sector work carried out by the World Bank and its member governments in the Africa Region. It is published periodically by the Results and Learning Unit on behalf of the Region. The views expressed in Findings are those of the author/s and should not be attributed to the World Bank Group. Malawi: Some Targeting Methods in Public Works Programs Malawi is among the five poor- cessfully reached the poorest. Ben- est countries in the world in des- efits have often been concentrated perate need of safety nets inter- on those with close ties to commu- vention. Frequents shocks such as nity leaders, as revealed by the recurrent droughts, HIV/AIDS, ris- Targeted Input Programme (TIP) ing orphanhood and widowhood administered by the government have all contributed to rising pov- which used Village Task Forces. erty levels in the country. It is in Communities had argued that this context that the country has they were all poor and in need of implemented safety nets assistance, hence they were un- programmes; given that the ex- willing to differentiate themselves treme poor1comprise 28.7% of the within various wealth categories total population while the poor2 in employed in the CBT. Further- general are estimated at 65.3%. more, the indicators used for tar- Given this widespread levels of pov- geting were too general to identify erty, the application of appropriate the poorest deserving individuals; targeting mechanisms has been and community leaders were not critical in the main safety nets trained on how to carry out the se- programs (public works, direct lection process and had no incen- transfers, and nutrition interven- tive to do this work with a sense of tions). responsibility. Community Based Targeting Considering these problems, (CBT) and Self-targeting are the CARE Malawi decided to use CBT two main methods used in the larg- from a different approach; with est programme of Public Works. community members selecting This note assesses and compares the these people using a set of criteria effectiveness and efficiency of these which they developed. The selected targeting methods in identifying the people were not only accountable poor in Public Works Programmes funded by MASAF and CARE in 1. There is no specific definition from literature. However Malawi's Central Region. It further in this research these people are defined as those who Findings seeks to identify challenges that hinder the effectiveness of these cannot meet their food requirements throughout the year methods3. from their own harvest. 2. Those who can meet their food requirements for at most 3 Chiefs and other community leaders/representatives have been - 4 months in the year from their own harvest used in the administration of CBT, 3. Based on MASAF sponsored research work conducted but these methods have not suc- by Mrs. Martina Lembani to the community, but they were portion of non-poor people who par- most appropriate for the different trained and supported to carry out ticipate in the programme out of programmes depending on the ob- this work. Transport and a monthly the total number of participants jectives of the programme. honorarium were provided as in- while under-coverage rate is the centives. CARE was working on a proportion of poor people who do not Table 1: Leakage and under-coverage rates by targeting method and level of pilot; able to conduct meetings in participate in the programme out analysis. all communities to select facilita- of the total population of the poor. tors, train them and pay them. A country- wide program would re- Data Analysis quire significant resources, care- Data analysis was done at two ful planning, coordination and su- levels. This was because the data pervision ­ all beyond the current collected contained three catego- capacities and resources. The ries of people (very poor, poor, and CARE CTB method is in contrast to better off), but the calculation of the self-targeting method used by leakage and under-coverage re- District Assemblies who employ a At both levels of analysis, the Z quired only two groups of people to foreman to register beneficiaries test results show that there is a be compared (poor and non-poor). on a `first come first save' basis in significance difference in the leak- Level one is where the very poor MASAF Local Authority Managed ages rates for the two methods at were considered as one category Programmes (LAMP). 95% significant level. This there- while the poor and better off were As in other safety nets fore leads to a conclusion that CBT combined to form another category. programmes, the choice of a target- is more effective at targeting the The justification for this is that ing method has been an issue, and poor compared to Self-targeting safety nets in Malawi are supposed this research was carried out to because CBT experienced lower to reach out to those within the guide the implementation of future leakage rates than Self-targeting 30% poorest population who are programs. The study compared CBT (i.e. 17% and 50.8% at level one capable of moving out of poverty. with and Self-targeting methods analysis and 2.6% and 13% at level The very poor are therefore covered used under the Improving Liveli- two analysis respectively). in this figure of 30% since the ex- hoods Through Public works treme poor in Malawi are esti- Programme (ILTPWP) managed by mated at 28.7%. This analysis CARE Malawi (which used CBT in The two methods are more effective at targeting the poor in six districts in the central region general thanthan the extreme poor of Malawi) and the MASAF-sup- who are the target group for ported nation-wide PWPs managed safety nets. CBT is more effective at targeting by Local Authorities. In this study, the poor compared to the Self- targeting method. food security was used as indicator As for under-coverage rates, the Z for CBT to identify the wealth sta- test reveals that there is no sig- tus of people after careful consul- nificant difference in the two tation with the community. helps to draw policy implications on methods in their under-coverage safety nets targeting in future. rates. This can also be verified by Level two is where the very poor the small differences that were and poor categories were combined, observed in the two methods at Measuring Effectiveness of the and the better-off were considered both levels of analysis. CBT had an Targeting Methods as a category on their own. This under-coverage rate of 69.6% while assessment reveals outcomes of Effectiveness of the targeting Self-targeting had 65.3% at level the general pattern of the poor ver- methods was measured by using one analysis. At level two analysis, sus the non-poor. This analysis leakage and under-coverage rates. CBT had 72.9% while Self-target- helps programme designers to de- Leakage rate is defined as the pro- ing had 66.2% as indicated in Table cide on which targeting method is 1 above. On this basis, none of MK959,510.00 (US$8,722)4 used to into different wealth categories the methods can be considered train District Assembly staff and as they insist that they are all more effective than the other. How- Community Based Facilitators poor. Communities do this to ever, these results indicate that (CBFs) who were used for facilitat- avoid conflicts, and it takes ef- both methods are not able to reach ing the targeting process in the fort and good facilitation skills to out to most of the poor people due communities, as well as payment manage this process. Proper to the limited resources, which of monthly honorarium and trans- training of the facilitators to permit fewer numbers of people to port allowance to CBFs. handle the process can help. benefit from the programmes. This On the other hand, the District 3) the process is sometimes taken simply confirms the extent of pov- Assembly staff who use Self-target- to family level from village level erty in Malawi that it is indeed ing methods on the LAMP could not because only a few people are al- widespread and deep. estimate the amount of money lowed to participate in the spent on registration of beneficia- programme because of the high Does it matter which targeting ries. They indicated that they demand for Public Works. A com- method is used to reach beneficia- train a Project Implementation munity decides to take one per- ries? Committee which is supposed to son from each family, which dis- The results of the survey further work in liaison with the foreman advantages poor families and ad- indicated that there is a higher on various project issues includ- vantages the better off ones. (almost twice) chance for the poor ing registration of beneficiaries. to participate in PWPs in commu- No estimate could be made on how Challenges in Self-targeting in- nities where CBT is used compared much of this cost can be attributed cluded: to where Self-targeting is used. to beneficiary registration. As 1) poor communication on benefi- Second, the results also revealed such, it is not possible to compare ciary registration, and some that at level two analysis (where the two methods in terms of effi- people only learn about the the very poor and poor are com- ciency due to lack of data. However, projects after registration is over. bined) there is an even higher discussions with staff from CARE 2) corruption can occur because chance of the poor participating in who used CBT revealed that they registration is done by a single the programme compared to level felt that their targeting approach foremen, who sometimes brings one (a situation where the very was costly. On the other hand, the people from places other than poor are a category on its own). district staff felt that their approach where the PWP is taking place This therefore shows that the two does not cost them much. (observed in some of the places methods are more effective at tar- that were surveyed). Factors influencing ineffectiveness geting the poor in general than the of the targeting methods Administrative issues of non- extreme poor who are the target payment or late payment as well group for safety nets. This has im- The research established that low wages affected targeting in plications for safety nets objec- there are some challenges in ad- some places because some of the tives. ministering beneficiary selection poor considered the wages exploit- using these methods. In terms of ative. Measuring Efficiency of the Target- CBT it was established that:- In Focus Group Discussions ing methods 1) it requires much time to under- (FGDs), communities were more Administrative costs were used take a thorough selection pro- positive towards the CBT method to compare the efficiency of the two cess because all community compared to Self-targeting; and targeting methods. The informa- members have to be present dur- opted for CBT in future programs. tion that was collected revealed ing the selection process to dis- that the CBT that was used by cuss and agree on who deserves CARE in ILTPWP took up approxi- to benefit. PRA tools are used to mately 0.1% of the total programme ensure transparency. budget on targeting alone. The 2) communities are not always 4. Used exchange rate of US$1 = MK 125 actual figure was calculated to be willing to categorize themselves Conclusion and Recommendations standing, it is very important for reached with safety nets Public development practitioners to con- Works Programs. The research revealed that effec- sider the objectives of the tiveness of the targeting method programme before making a choice This article was written by Martina partly depends on how much atten- on which targeting methods to be Lembani and Charles Mandala as a tion and effort is put into the pro- used. In the case of safety nets MASAF Policy Infobrief, Vol. 1, Is- cess. The case of CARE's CBT ap- programmes, it is appropriate to sue1, 2006. A detailed discussion of proach shows that the effort they use the CBT method, which effec- the issues contained in this Brief will had put in the targeting process tively identifies the extreme poor be found in MASAF Discussion Pa- achieved the desired results of tar- compared to Self-targeting. How- per No. 01/2006; Assessing the geting safety nets, which require ever in a situation where the num- Effective and Efficiency of Target- the poorest to participate in the ber of poor persons is very large and ing Methods in Public Works programmes. In addition, other pre- the programme targets the poor in Programmes in Malawi: the Case conditions such as administration general, Self-targeting is still an of MASAF and CARE Managed of payment and wages also have to option. Another factor that has to Programmes in the Central re- be considered to ensure effective- be considered is the scale of opera- gion of Malawi. ness of targeting the poor. In tion. Programmes that concen- terms of efficiency, which concen- trate on small geographical area trated on administrative costs of For further information, please like the CARE and other NGOs are the programmes, the research did contact: able to undertake the detailed CBT not get enough data to make a con- approach due to their setting up MASAF Management Unit, Red clusion on the two targeting meth- institutions to implement and Cross House, Private Bag 352, ods. However, based on discussions monitor the activities. In a nation- Lilongwe 3, Malawi. Tel: (265) 1775 with staff from CARE and District wide programme, appropriate local 666/174/702; Fax: (265) 1775 949 Assemblies, it was concluded that institutions could be identified in CBT was more costly compared to order to reduce administrative Self-targeting. With this under- costs so that the poorest can be