''"…there are no incommensurables when decisions are made in the real world'' (S&Z 153.5).

Basic Concept and Fundamental Rule

Efficiency : Put resources to their best possible use. Select the alternative with the greatest net benefit.

Generic Scenarios and Corresponding "Rule"

  1. '''Yes or no on a single project''' : If the net benefits are greater than zero, the project is worth doing
  2. '''Pick one of several options''' : select option with largest net benefit.
  3. '''Choosing scale for a project''' : Stop expanding project when marginal benefit balances marginal cost.
  4. '''Accept or reject multiple projects subject to resource constraint''' : sort by "bang for buck" and calculate cumulative cost up to limit. If lumpy goods and cost goes over, swap out last and bring in next affordable option.
  5. '''Accept or reject several and choose scale''' : Distribute resources so marginal benefit is same across projects..

Heuristics (after Weimer and Vining)

  1. Don't expect to find a perfect alternative.
  2. Don't compare a preferred alternative with fake alternatives.
  3. Don't pick a favorite before you have looked at them all.
  4. Make alternatives mutually exclusive.
  5. Avoid "do everything" as an option.
  6. Make alternatives consistent with available resources, authority, etc.

Generic Concepts

  1. Pareto improving: a change that makes some better off but no one worse off.
  2. Kaldor-Hicks criterion: adopt a policy only if those who will gain could fully compensate those who would lose and still be better off.
  3. Opportunity cost. Value of required resources in their best alternative use.
  4. Willingness to pay.


  1. Identify all stakeholders
  2. Identify costs and benefits and convert to common units
  3. Determine what type of problem — number of alternatives, scale, budget constraints
  4. Calculate net benefits and marginal net benefits
  5. Analyze and choose

Bottom Lines

  1. Cost-Benefit Rationale (see [WWW]http://www.stanford.edu/class/msande290/290_05_Cost_Effectiveness.pdf)
  2. A policy change is a Pareto improvement if some people are better off and no one is worse off after the change.
  3. Some policy changes benefit some at the cost of others. An exchange could have those who benefit compensate those who suffer, and thus make everyone better off.
  4. A policy change is a potential Pareto improvement if an exchange could be made among people that would make it a Pareto improvement, even if that exchange never occurs.
  5. A policy change is considered desirable if it is a real or potential Pareto improvement. This is determined by accumulating its direct and indirect benefits and costs.

A Note on Cost Effectiveness

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.

  1. 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.
  2. "Bang for your buck"
  3. Useful when you cannot (or do not want to) express costs and benefits in a common metric.
  4. Does NOT allow us to figure out whether benefit justifies price or to select levels appropriate for a project.


  1. Costs of alternatives are same so we only need to compare benefits
  2. Benefits of alternatives are same so we only need to compare cost

Avoid Benefit/Cost Ratio

Suppose I can announce an new initiative (at the cost of $1 for the time it took me to tweet it) and this will produce $100 of public welfare — people will just feel good knowing I care. Or, I can invest a day of work (at my rate, that's, say, $200) and produce $1000 of public welfare. The benefit/cost ratio of the first is 100, the second just 5. But, if I'm not constrained in my budget, the second is obviously better for the world.


  1. Inputs to CBA require work (models, calculations, research)
    1. Identifying all costs and benefits tricky (second order, opportunity, side effects, unintended consequences)
      1. Valuation : market prices, willingness to pay
      2. Costs: opportunity, intangibles
  2. Important to look at marginal rates — a little more, a little less, what's going on "at the margins"?
  3. Redistribution. In essence, CBA sets these aside which is NOT the same as devaluing them.


Benefit-Cost Ratio
total benefits divided by total costs with criteria of proceed if greater than one, reject otherwise. Logic is usually to accept project if it is > 1 and reject if < 1 or adopt the alternative with highest ratio. SOMETIMES this gives same result as "maximize net benefits" but NOT ALWAYS. Can produce results that are contrary to the maximize net benefit principle. Problematic with mutually exclusive choices and resource constraints. Sensitive to how costs and benefits are tallied (see Stokey & Zeckhauser 146.6) In general, DON'T USE.

Example. Imagine two options. One is to spend $1000 on advertising with expectation of $10,000 in benefits. The other is to spend 25,000 with expectation of $100,000 in benefit. The first option gives you a better "bang for the buck rate" (i.e., benefit/cost ratio) but provides 90,000 less in benefits. If we don't have a resource/budget constraint it is the better choice .

Cost Effectiveness
a truncated version of CBA in which either all the benefits or all the costs are identical allowing comparison to be made only on the other measure. (google, wikipedia)
putting resources to their most valuable use
ex ante
def. from the Latin for "before the event"; refers to any kind of analysis done prospectively — before we know the outcomes. The opposite of ex-ante is ex post. (See also google, wikipedia)
Fundamental Rule Of CBA
Select the alternative with the greatest net benefit.
Kaldor-Hicks Criteria
a changes is warranted if those who gain could compensate those who lose and everyone would still be better off. Separates distributional (equity) concern from efficiency concern.
Law of diminishing returns
the idea that most interventions eventually produce less and less of their desired outcomes. Or, that most outcomes are not forever linear in their inputs.
Opportunity Cost
the value of other uses to which a resource might have been put
Pareto Criteria
a change is only warranted if it makes someone better off without making anyone worse off
Willingness to Pay
In cost benefit analysis, the amount a stake-holder would pay to have a project go forward if her contribution would make the difference between doing and not doing the project. Useful as a way of accounting for interests of all stakeholders, even those too poorly organized to participate in negotiations.