Some time ago a question was submitted to the “Answer Desk” section of this blog on what the steps in cost-benefit analysis (CBA) are. Let me try to take a belated and brief crack at this, with the caveat that entire volumes have been written on the subject. I will primarily try to give some insight on what the steps are in theory and why they are difficult in practice, and close with a comment on how CBA is conducted in practice.
In theory CBA is easy. In evaluating any environmental action (or other public intervention) we need only list out the potential costs and benefits of the action, assign monetary measures for each category in the list, add them up and compare them, and if the benefits are greater than the costs the action is ‘worth it’ from an efficiency perspective. In practice each of these steps is fraught with difficulty. In some respects qualitatively listing the categories of benefits and costs is the easiest part. For example, there was a post discussing the benefits and costs of dam removal from western streams. The benefits categories of this can include improved stream ecology, enhanced wildlife populations, and better moving water recreation opportunities. The costs include loss of electricity generation, flood control, and flat water recreation. These lists, however, are almost certainly incomplete. Most policy actions have some degree of uncertainty and unintended consequences. In listing the categories practitioners try to focus on the first order effects, while acknowledging ex ante uncertainty.
Assigning monetary values to the categories is the most difficult and controversial of the steps in CBA. It is necessary, however, for comparison of benefit and cost categories that are not expressed in the same units. Since many (often most) of the benefit and cost categories are non-market – which is to say they are not bought and sold at market prices – an extra step is needed to assess the categories’ dollar values indirectly. For this the sub-discipline of non-market valuation is used. Research in non-market valuation has developed fairly sophisticated theoretical and empirical methods for describing and measuring non-market values. Empirical research has examined the value of flat water vs. moving water recreation, the existence of particular wildlife populations, and benefits of ecosystem services such as flood control to name but a few. Estimates of these types of environmental services are the building blocks needed for assigning monetary values to non-market services.
Once the benefits and costs are measured in dollar terms an additional difficulty must be addressed. Typically benefits and costs do not occur at the same time. Again referring to the dam removal example, electricity generation will be lost immediately but wildlife populations may take years to recover. In this case the cost occurs right now, but the benefit is not available until later. Compare this to a monetary transaction. A hundred dollars awarded today is not the same as a hundred dollars awarded two years from today. To compare the two, adjustment for the interest potential is necessary. This is similarly needed for CBA. If we want to compare benefits and costs occurring at different time scales discounting is needed to express future costs or benefits at today’s equivalent value. Discounting is mechanically easy, but no agreement exists on what the correct discount rate is. Controversy over discounting lies at the heart of the debate on CBA, in that the choice of discount rate can often determine whether net benefits are found to be positive or negative.
Once each challenge in the CBA process has been addressed and the discounted costs and benefits calculated there remains the challenge of interpretation. CBA uses a very narrow criterion for determining the desirability of an action. The focus is on efficiency, and nothing is said about distribution or fairness: from a pure theory perspective an action would pass a CBA if all the benefits accrued to one wealthy person and the costs were born by the most disadvantaged members of society. Most would agree this would not be a desirable outcome. Thus CBA should not be the sole determinant of a decision, but rather one of several used to summarize available information.
So how is CBA done in practice day to day? My experience here is a bit more limited, but I will make a few observations and invite comments from those more experienced. My sense is that there is a large reliance on ‘off the shelf’ benefit and cost estimates rather than original studies due to time and money constraints. My sense is also that much of the work is qualified: since it is usually impossible within time deadlines to account for all potential costs and benefits analysts provide monetary estimates for categories that can be determined and acknowledge categories that can not. Thus a full accounting of even the efficiency properties of an action is usually beyond reach. This said I still find the process quite valuable. As I’ve argued elsewhere, the structure and transparency the CBA process provides when conducted carefully can organize and help discipline public decision making.
