A cost-effectiveness analysis (CEA) compares the cost and effectiveness per unit of a given program to determine whether the value of an intervention justifies its cost. CEA provides the metrics to rank or compare similar interventions or projects that result in the same effect.
- A CEA should be included in the project design in order to compare alternative projects.
- While a CEA calculates the cost per unit of effect, a CBA calculates the ratio of all costs to all benefits of a program.
- CEAs require careful calculation. Since CEAs are sensitive to place, scale, and errors in estimates, it may be difficult to precisely compare programs.
A CEA measures the cost per unit of effect (i.e. the cost to increase skilled birth attendance by 50 percent) or visa versa (i.e the percentage gain in skilled birth attendance induced by spending $100). The analysis must accurately reflect the program based on costs and impacts actually observed.
A CEA is useful for policymakers looking to compare programs when they are primarily concerned about one outcome of interest (i.e. increasing skilled birth attendance rather than maternal and child health broadly). It can be used to compare multiple interventions with common outcome(s) if the costs and benefits are computed using similar methodology for all programs.
Cost-effectiveness analysis (CEA) vs. cost-benefit analysis (CBA)
CBA calculates the monetary ratio of all costs to all benefits of a program. CBA can help to determine whether a program is worth the investment. It can also allow comparison across vastly different interventions (i.e. education versus agriculture). However, CBA requires a number of strong assumptions about the monetary value of all the different benefits, including the lifetime benefits of an intervention. In contrast, CEA is a transparent, simple, and objective measurement that enables comparison of programs with common outcome(s) of interest. However, the implicit assumption is a common post-intervention trajectory.
|CBA Formula||CEA Formula|
Considerations for cost calculations
- Use program financial records. Typically, begin calculating costs by referring to program budgets, which provide a list of all relevant activities involved in implementing the program. Consider, however, that budgets are forward-looking estimates of true costs. Thus, it’s better to use actual program financial records on expenditures for calculations. However, this is not commonly done.
- Interview program and field staff to get key information including unit cost data like allocation of staff time across activities and wages for staff at various levels of implementation team. Note that using average wages makes calculations less sensitive.
- Model program costs at the margin. You want to capture marginal costs of new activities initiated as part of the interventions. For example, to calculate the costs of a school meals program, you may need the cost of new kitchen, cooks, implements, and food. However, you should not include the cost of school administration or buildings as these would exist whether or not the intervention was implemented.
- Consider depreciation. Account for how to value new assets or equipment obtained over the program implementation period.
- Consider the cost of user time. Include the costs of participation in program (i.e. using local wages).
- Differentiate between pilot costs and scale-up costs.
- Include spillover effects on indirect beneficiaries when quantifying impacts
Challenges to conducting CEA
- It is difficult to get data on program costs, especially from other authors. This is where Field Coordinators come in, as they have an advantage in collecting local costs.
- Since CEAs are sensitive to place, scale, and errors in estimates, it is difficult to compare studies. Exchange rates, inflation rates, discount rates, differing factor endowments, and the efficiency of implementors all affect CEA results and make valid comparisons between studies or programs difficult.
- Because CEA measures cost of obtaining impacts on a single outcome, aggregating cost-effectiveness across multiple outcomes is difficult
Use in impact evaluations
Very few papers have undertaken comparative cost‐effectiveness of different programs. A meta-analysis of randomized experiments finds that of the 77 RCTs gathered, 56% don’t report any data on incremental costs. Studies that have incorporated CEA include Kremer, Miguel and Thornton’s Incentives to Learn and Banerjee, Cole, Duflo, and Linden’s Remedying Education.
Many researchers have no incentive to conduct a CEA. There is sometimes a trade-off between exposing program effectiveness and cost-effectiveness. For those that do conduct CEAs, it is unclear how to report insignificant results on outcomes. It is important to think about how insignificant CEA results should be interpreted and presented in such cases.
Back to Parent
This article is part of the topic Cost-effectiveness Analysis
- JPAL's case study on estimating the cost effectiveness of education programs.
- JPAL's slides on CEA and scaling up
- IADB's Cost Benefit and Cost-Effectiveness page
- McEwan's Cost-Effectiveness Analysis of Education & Health Interventions in Developing Countries
- Dhaliwal et al.'s Comparative Cost-Effectiveness Analysis to Inform Policy in Developing Countries
- Evans and Popova's Cost-Effectiveness Measurement in Development: Accounting for Local Costs & Noisy Impacts
- Cost-Effectiveness Analysis: Methods & Applications, by McEwan & Levin (2000)