- Stokey & Zeckhauser. 1977. "Project Evaluation: Benefit-Cost Analysis"
- Download worksheet for examples.
Benefit/cost analysis has, perhaps more than any other analytical technique covered in this course, become an everyday concept. And, fortunately, the real thing is, more or less, the same as its popularization. But even if the concept is well known, how to carry out a cost-benefit analysis well requires a little training.
Definition: The systematic enumeration of all tangible and intangible benefits and all costs, whether easy or difficult to measure, to all members of society if a particular project is undertaken.
This approach is generally intended to be ex-ante – that is, we will attempt to evaluate projects before they are implemented. We lay out how to deal with multi-attributed outcomes, set up and carry out benefit/cost and cost effectiveness analysis and zero in on when they are appropriate and what limitations they are subject to.
How We Will Proceed
- Two examples
- Describe basic formal procedure
- Explicate some components of this procedure
- Introduce variations
- Examine common mistakes and misconceptions.
- Review indications, contra-indications, caveats and practical advice
Introduction: Two Examples
The Standard Canard
Imagine you are a company manufacturing a consumer product1 of which you sell 100,000 per year. A fixable flaw is discovered. You determine it would cost $13 per unit to eliminate the flaw. You also determine that the chances of a failure in the product that would result in a $10 million dollar law suit are 1 in 100. How do you think about your options?
REVIEW: Calculate the net cost/benefit for each of two strategies: (a) fix the problem, (b) ignore the problem.
What is the obvious flaw(s) here?
Consider the start of the semester. Our lab was scheduled for Tuesday and I discovered that it would be much better for me if the lab were scheduled in the same time slot but on Thursdays instead. How should I think about addressing this problem?
- Perhaps I am in charge and I get to decide and so I just make the change by fiat.
- Or, I think that students have already made plans around the schedule and so it is simply too late: I must eat the cost of the the inconvenient schedule.
- Or, I note that course time slots are Tu-Th so it is very unlikely that anyone in the lab can be taking a class that meets on Thursdays in the time slot and so I make the change by fiat.
- Or I follow the same theory but verify it empirically by surveying the class: "Does anyone have another class that meets Thursday 1-230?"
- Or I recognize that there could be various reasons that the change might be an inconvenience for students who'd made plans on the basis of a Tuesday class. An so I embark on a more formal cost benefit analysis to determine which alternative is "better for the world."
Work through some of the logic I might employ if I pursued the last option above.
What have we learned in these two examples?
The first example gives us pause, perhaps even suggesting that cost benefit analysis is immoral — are they not, after all, putting an arbitrary price on human life?
That example also reminds us that we will use the tool of expected value as an INPUT to cost/benefit analysis.
The second example introduced four other concepts — the search for stakeholders, net benefit, "willingness to pay" as a measure of value, and compensation "in principle" as a logic of policy choice. Each of these needs to be examined in more detail. We'll do a little of that here as a preliminary with more examination along the way.
Stakeholders. When we do benefit-cost analysis as policy makers we are assessing the big picture. Our first task is to make some decisions as to how we will define "the whole world" of individuals, communities, and organizations that might be affected by the things under consideration. We are especially mindful of What we have in other contexts called externalities.
Net Benefit. "Netting" refers to adding up the offsetting pluses and minuses to reach an overall measure of benefit provided by a project. We do this for each stakeholder and for the "community" of stakeholders as a whole.
Willingness to pay. In our quick and dirty analysis above we resorted to asking the question, "what would so-and-so be willing to pay for X?" as a way of figuring out what the benefit of X is for so-and-so. This move may feel a little artificial, but it is a tried and true convention for this type of analysis.
Compensation in principle. This is perhaps the hardest idea for beginners to get their head around — and it's subject to ongoing debate among scholars and experts. The point to grasp here is that the question of WHETHER actors compensate one another is separated out as a political question of redistriubtion. Basic benefit-cost analysis points us toward the biggest pie; it leaves for other forms of analysis the question of how that pie is divided.
So what did we do?
How does CBA fit into the general schema?
The basic process in CBA is always the same:
but the details may differ depending on the circumstances. There are three questions that combine to define what we mean by circumstances. They are
- Are we thinking about one project or several alternatives?
- Are the projects of fixed size (scale) or do we have to decide on how large or small to implement them?
- Are we subject to a budget constraint?
|Category||More than one||Scaleable||Budget Constraint|
|A||No||No||No||Simple question: should we do this or do nothing? (see case 1)|
|B||No||No||Yes||Should we do this and can we afford to do this? (see case 1)|
|C||No||Yes||No||When is a little more of the project not worth doing? (see case 2b)|
|D||No||Yes||Yes||What is optimal size of project? Can we afford to do that? (see case 2b)|
|E||Yes||No||No||Which projects should be done? (see case 2a)|
|F||Yes||No||Yes||Which projects should we fit into our budget? (see case 3)|
|G||Yes||Yes||No||Which projects should be done at what scale? (see case 4)|
|H||Yes||Yes||Yes||Which projects should we do, how intensely, when do we stop? (see case 4)|
STOP AND THINK. Sketch a flow chart with the above criteria as cascading decision diamonds leading up to recognize that we have a Case 1, 2a, 2b, 3, or 4 type CBA problem.
Case 1: Is a given project worth doing?
Case 2a: Choosing one project among many
Case 2b: Choosing right scale for a project
Case 3: Multiple Projects with Resource Constraint
Case 4: CBA Multiple Scalable Projects with Resource Constraint
- Garbage in Garbage out. Bad assumptions can distort analysis.
- Who ARE the stakeholders?2
- Externalities. Easy to omit side-effects.
- Easy to forget to pay attention to what's happening at the margin — we focus on the total and don't notice that a little more investment might provide lots of benefit.
- Estimation of costs and benefits is hard.
- Intangibles hard to value
- Opportunity costs distorted by availability
- Opportunity costs easy to miss.
- Claimed willingness-to-pay may be unreliable
- Mixed (high/low) mis-estimates can totally muddy output. Important to slant all assumptions in same direction.
- Distributional effects DO matter, but in its traditional form, benefit-cost analysis does not address distributional issues.
- False impressions of precision and objectivity
Best Practices to Meet Challenges
- Explore unexpected side-effects
- Look for sensitive dependence on assumptions
- Structure process so that finer analysis can be performed as better information becomes available.
- Use willingness-to-pay, if even only as a thought experiment
- Defer pricing intangibles until late in process.
CBA has costs and benefits in same units — that's why we can subtract and get to net benefits. Sometimes getting them into same unit form is either not possible or undesirable.
- Sometimes the benefit is fixed – we need a way to feed 2000 homeless people this weekend – and sometimes the cost is fixed – we have $45,000 to get word out about this new vaccine.
- "Bang for your buck"
- Useful when you cannot (or do not want to) express costs and benefits in a common metric.
- Does NOT allow us to figure out whether benefit justifies price or to select levels appropriate for a project.
- Costs of alternatives are same so we only need to compare benefits
- Benefits of alternatives are same so we only need to compare cost
- AN ALGORITHM FOR COST-BENEFIT ANALYSIS
- Wikipedia on Cost Benefit Analyis
- “Cost-Benefit Analysis: An Ethical Critique (with replies).” by Steven Kelman, from AEI Journal on Government and Society Regulation (January/February 1981) PP. 33—40. Reprinted with permission of the American Enterprise Institute for Public Policy Research, Washington, D.C.
- New York Times: "Harvard's Kennedy School: Is Competence Enough?" By J. Anthony Lukas: J. Anthony Lukas. Published: Sunday, March 12, 1989.
- SYLLABUS, API-102B/C/D: ECONOMIC ANALYSIS OF PUBLIC POLICY, JFK School.
- An Illustrative Hypothetical Example of Cost-Benefit Analysis: The Parking Problem