Together with staff from the Africa region, DDVE recently completed an evaluation of the targeting performance and impact of a Bank funded cash for work program in Liberia. The program was implemented within the context of a strategic priority placed by the President and the Government of Liberia on youth employment. The Cash for Work Temporary Employment Project (CfWTEP) was financed by the World Bank Food Price Crisis Response initiative, and the evaluation of the project was conducted using a light survey instrument (at a data collection cost below $25,000) with four objectives: (i) Assessing the targeting performances of the program; (ii) Measuring the wage substitution effects among the participant; (iii) Analyzing the patterns of use of the wages received by households; and (iv) Documenting other aspects of the program.
The basic principle for the evaluation consisted on assessing the welfare status of participants through a matching procedure whereby the survey data on program participants was appended with the nationally representative 2007 CWIQ survey, and the two data sets were then used together for matching treated program participants with an appropriate control group. Because the questionnaire of the CfWTEP survey was designed in such a way as to be comparable to that of the CWIQ survey, a large number of variables could be used for the matching procedure. Both traditional propensity score matching and a combination of coarse exact matching and propensity score matching were used for the estimations of the level of consumption of the households to which program participants belong. The results suggests that about three in four programs participants could be considered as poor, suggesting rather good targeting, even though the likelihood to freach the extreme poor was a bit lower.
A series of questions in the evaluation survey were used to assess the income gains of program participants thanks to the program. Wage losses for program participants may arise due to the fact that they may have had another paying job before the program, or that they may have needed to hire a worker to replace them in their previous occupation or to help for domestic work while they participated in the programs. Yet these losses tended to be very small as compared to the incomes gained from program participants. Thus, wage substitution effects did not seem to play an important role in reducing the effectiveness of the program in generating higher additional income for participants, probably in part because there is a large mass of very low paid or unpaid workers in Liberia, so that those who take up a job with CfWTEP are unlikely to have had gainful employment as an alternative to program participation.
The estimates of the wage gains were then used to assess poverty levels before and after program participation among participants, simply by considering that any additional wage gain or loss translates into an equivalent change in consumption level. The findings of the assessment suggest a substantial reduction in poverty among program participants. A series of additional questions on subjective perceptions regarding the program were also asked to program participants, suggesting favorable ratings in many dimensions such as whether program participants had gained technical skills, learned to be punctual, considered the wage paid as fair, and benefited from the first time from a bank account (wages were paid through such bank accounts). The results of the survey were also encouraging regarding the potential longer-term impact of the program on households through their expenditure choices for the additional income gained. As much as 30 percent of the income of participants was spent on education (this may have been influenced by the timing of the program, where most wages were paid when school fees were dues) and 25 percent on various types of investments. This suggests that the program may not only be achieving short-term impacts, but that it may also have some longer-term impacts on the targeted households.
This work benefitted from the funding of the Gender Action Plan. For questions on this analysis, please contact Quentin Wodon at firstname.lastname@example.org