CEA compares the cost per unit of effect in a particular project or program option with the costs per unit of effects for alternatives. The comparison between costs and effectiveness will allow the ranking of the alternatives or a comparison with similar interventions or projects. It is important to underline that cost effectiveness measures allow for the ranking of interventions that result in the same effect.
Linkage between objetives and measures of effectiveness
The measure of effectiveness (and cost effectiveness) should be directly linked with the stated objective of the project. In any project, there are many possibilities as to which measure of effectiveness to choose from. Effectiveness selection should be guided by two criteria: reliability and validity.
- An indicator is said to be reliable if it yields similar results when applied repeatedly on the same individuals or populations. As most measures of effectiveness do not have entirely reliable indicators, it is suggested that the cost effectiveness analysis includes a reasonable range for these effects.
- An indicator is said to be valid if it has a close match to the underlying impact it is trying to capture.
When the analysis uses intermediate outcomes that have limited validity, it is suggested that the analysis – and justification – for the selected indicator be based on empirical evidence from secondary sources.
Level at which cost effectiveness should be measured
Effectiveness can be measured at both outcome and impact levels. In some cases it can be necessary to resort to intermediate outcomes, if relevant outcome or impact information is scarce. If intermediate outcomes are used, evidence should be presented on the relationship and correspondence of these intermediate outcomes to the expected impact, if the objective is stated in those terms.
Satisfying any particular need (clean water, for example), or solving any particular challenge (reducing the crime rate, for instance) can be done in a myriad of ways. Alternatives should always be specified as to addressing the same problem, solving the same challenge.
Alternatives share the problem and differ as to the solution. Alternatives can only be compared when addressing the same or similar challenge. As any problem can be addressed in many ways, it should be taken into account that the number of alternatives considered does have a cost and time implication and should be limited. Alternatives typically involve choices on scales, degrees of involvement (doing nothing, doing nothing plus, upgrading), timing in investing, institutional designs and technological choice.
Once the cost effectiveness ratios have been calculated alternatives need to be assessed based on the calculated ratios. Alternatives should be ranked according to their cost effectiveness ratios.
All of the considered alternative approaches should have cost effectiveness ratios that reflect the same impact, and the alternative with the lowest ratio should be selected. If more than one impact is included, particular care should be placed in cost allocation to each; all costs and effects should be expressed in marginal or incremental terms.
There are different approaches in identifying costs including the ingredient method which relies on the identification of all resources or ingredients consumed in an intervention and the valuation of each ingredient. The specification of ingredients is often facilitated by dividing ingredients into categories including 1) personnel, 2) facilities, 3) equipment and materials, 4) other program inputs, and 5) beneficiary or client inputs.
For Bank-funded projects, public sector costs are identified in project documents and the ingredient method is most useful in identifying other costs such as operation and maintenance costs and non-public sector societal costs.
Note: Sunk Costs
Identifying costs in the project
Here are some of the costs the project must identify:
First: public sector costs included for the project investments
Project documents lay out investments and non-recurring costs explicitly as part of project preparation and there should be consistency between what is noted in the project and what is identified as project costs, including costs that are part of the loan as well as counterpart costs.
Second: additional operational and maintenance costs or other recurring costs to the public sector
If the project establishes new government entities or invests in new infrastructure that requires funds for operation and maintenance or other recurring costs beyond the project timeframe these should be included. These are incremental costs beyond what would have been paid for in operation and maintenance without the project.
Third: costs to society that should be incorporated - non-governmental organizations, firms, households or individuals
Any additional costs to non-public entities that result from the project should be included. In general, all inputs (costs) that contribute to the generation of benefits should be included. Otherwise benefits would have to be adjusted accordingly. These costs depend largely on the project and linked to the response required by project beneficiaries to obtain the benefit of the project.
Fourth: total flows of costs in the project
All types of costs should be noted and aggregated to get the total cost flows and the present value of those costs flows. It should be clear that all of these costs are incremental and the manner in which they are measured should be transparent. In the case of shared or common costs (with other projects or among components), these should be distributed and assigned according to a clear allocation rule, which needs to be explicitly stated in the assumptions section.
Sensitivity analysis allows the testing, under different scenarios, of the strength and robustness of the assumptions, the data and the logic of the intervention that underlie the economic analysis and to acknowledge the underlying uncertainty of the expected results. Any forward looking estimation presumes behavioral assumptions, assumptions on the vertical logic, and on the causality links that allow the attribution of a particular benefit or cost to the goods and/or services delivered or used by the project.
In addition it is necessary to undertake sensitivity analysis to compensate for the natural tendency for project appraisers to be overly optimistic. Many project parameters are affected by optimism, overstating benefits and understating timings and costs, both capital and operational.
This content is dynamic