Impact Evaluation of the Labour Intensive Public Works (LIPW) in Ghana

Public works programmes (PWPs) have become an increasingly available policy option in tackling this socio-economic menace, and also as a social protection tool. This study analyses the impacts of Labour Intensive Public Works (LIPW) Programme on welfare in Ghana. This programme is being undertaken by the Ministry of Local Government, Decentralisation and Rural Development (MLGDRD) with funding from the World Bank. The LIPW programme provides seasonal paid employment to selected households/individuals located in project communities.

The LIPW offers temporary wage-earning opportunities to poor households during the agricultural off-season. In addition to providing wages, the program supports the creation, rehabilitation, and maintenance of public or community assets including rural feeder and access roads, basic water infrastructure (small earth dams and dugouts), and environmental degradation and climate change mitigation activities (reforestation and afforestation, catchment and watershed protection, and biodiversity conservation).

The target group for LIPW programmes is the poor able-bodied individuals who require work during the 'drought' or agric-off seasons. Public works programmes reach the target group through a variety of mechanisms and design features that include wage rate policy, self-targeting. targeting, regional targeting, seasonal targeting, and specific selection of households (for example, displaced households) and household members.

The targeting strategy for LIPW is to first geographic targeting of less developed districts based on poverty maps, followed by decentralized selection of prioritized communities within districts. Within the selected communities, the project relies on self-targeting of beneficiaries through wage rate setting ot ensure that only the targeted poor households volunteer for the public works. This may be supplemented by community-based targeting or rotation if needed.

Access to LIPW employment is expected to lead to improved labour market outcomes (employment and income) and welfare of poor households in the lean season (that is, when the LIPW project is active in their communities). Through greater physical and human capital accumulation (such as better health and nutrition) in the lean season, it could also lead to higher productivity (incomes) and consumption in the longer run. The opportunity to participate in LIPW could lead to female empowerment, reduction in migration, as well as better health and education of family members.

The public assets which will be improved through LIP Ware expected to have community-wide impacts. For instance, feeder roads could lead to availability of vehicles to convey goods to the market, shorter travel times and lower transport costs - which could improve the productivity of local enterprises, community plantations and their effect on food security and climate change adaptation and mitigations - while dams could reduce water stress related deaths of livestock, improve irrigation and hence farm yields. Further, climate change mitigation initiatives promote afforestation, and over time, contributing towards mitigating climate change. These community-wide impacts are expected to be reflected in improved household-level outcomes (such as income and consumption) among even those not directly employed in LIPW.

2024-06-01

2025-05-31

Paid employment, food security and nutrition status, health status, agricultural productivity, poverty status, seasonal migration, and overall well-being

Agriculture investment decisions, access to health insurance, health expenditure, education expenditure, women’s economic empowerment, food consumption expenditure, expenditure on productive assets.

The study relies on both quantitative and qualitative impact evaluation designs. Our primary objectives with this component are to establish the causal effects of LIPW on households and individuals, as well as beneficiary communities. Secondary objectives include an analysis of targeting effectiveness and assessment of the implementation performance of LIPW as reported by recipients. The core design will entail comparing a sample of LIPW households in 65 communities which will receive LIPW projects in 2024 (the intervention area – T) with households in comparable communities (i.e., 65 communities) which will receive LIPW jobs in 2025 (C1). An additional control group will be built from the treatment communities to measure the potential spillover effects by the project. This group, labelled (C2) will involve eligible households or individuals from the treatment communities who will not work for the project. The two main groups (T & C1) will be randomly assigned by the survey team based on project records. Thus, the evaluation is based on a randomized phase-in design, with randomization into phases performed at the community level. Impact will be evaluated by comparing outcomes at baseline and at follow-up. Treatment is expected to be staggered by at least one year. A baseline survey will establish the starting point for both samples, and subsequent follow-up in year 2 will allow a comparison of changes in the relevant indicators across the two study arms. Where resources are available we will recommend a light midline survey to enable the evaluation to minimize attrition and its effect on the outcome. It will also help us construct a panel and support the process evaluation. Since the LIPW intervention is implemented at the community-level, the primary sampling unit (PSU) or cluster is the community, of which we are proposing 130 in total. A power analysis indicated that 130 communities can provide adequate power within the total sample size indicated below, and still keep field costs manageable.

The proposed design can be described a phased-in Randomized Control Trial (RCT) approach supported by a mixed methods approach for the impact evaluation. We will use a difference-in-differences modelling approach to estimate program impacts for each outcome. The analysis will be conducted on the average treatment effect (ATE), the average treatment effect on the treated (ATT) and other econometric analysis where need be. ATE will be calculated for all relevant final outcomes of the intervention. A set of control variables will be selected to improve efficiency of the estimates. Multiple hypotheses testing based on randomization inference approach will be undertaken.

For the qualitative component of the evaluation, we propose an explanatory case study methodology with data collection at baseline and endline. By embedding a panel case-study approach into the larger quantitative component, we will be able to respond to all the evaluation aims of including: 1) LIPW Programme outcomes, 2) processes and outcomes of LIPW and complementary services on gender dynamics and inclusiveness; 3) household coping and resilience.

Not available

Randomization done in office by a computer

Community/sub-project

Yes

130 communities/sub-projects

2,691 households

65 treatment and 65 control