Benefit/Cost Analysis: Introduction

How do you choose between projects? You could always count "eenie-meenie-minie-moe," or flip a coin. But presumably you would prefer to choose between projects based on their merits, rather than chance.

The question is not as simple as it first seems. Any project is similar to any other project in some respect or other; every project is different from all other projects in most respects. The trick is to pay attention to those respects which will expose the most useful differences. Even then, every project is likely to be superior to all others in at least one significant respect. The idea is to find a way to compare those differences meaningfully.

The method most commonly used today for comparing projects is "Benefit/Cost Analysis," a method developed by the Army Corps of Engineers before World War II, adopted by business after the War, and making its way into local government use late in the 50's.

The basic idea of benefit/cost analysis is to select the project which gives "the biggest bang for the buck." To put it more formally, select the project that produces the greatest net benefit. Putting the issue this way sneaks in several assumptions and an armful of practical problems.

The first assumption is that one is dealing with roughly comparable projects. If projects are roughly comparable, then the only question is which achieves more of whatever you are interested in. If the projects are not comparable, comparing the net benefit from each project is beside the point: What sense does it make to compare a net benefit of "3" and "4," when one measures school lunches and the other, housing units? It may even be possible to convert these measures to common units (say, dollars); any comparison would still miss the point. Unless one can find some purpose that these projects have in common (alleviating poverty, perhaps?), comparing the results of the projects is futile. If the projects do have something in common, the results of the projects must be measured in those common terms (e.g., alleviating poverty) rather than the terms unique to each project (e.g., housing units and school lunches).

In formal discourse, this is referred to as the distinction between "effectiveness' and "efficiency." Effectiveness is a qualitative measure--is the goal achieved? Efficiency is a quantitative measure--how much of the goal is achieved at what cost. It makes no sense to talk about efficiency unless one is first satisfied that the projects in fact achieve what is desired. There is a story told that, when one of the astronauts was circling the globe, one of his buddies radioed up and asked him how he felt knowing that everything around him was "low bid." No one wants a low-bid product that doesn't work. Morris' article on "goals achievement matrices" (1968) is a particularly good discussion of this point.

The situation is made more complicated by the language that is often used. "Effectiveness" is commonly used as a shorthand for "cost-effectiveness," which is a measure of efficiency. Benefit/cost analysis itself is a tool for measuring efficiency. It is assumed that the analyst has guaranteed effectiveness by excluding projects which do not meet standards.

The phrase "net benefit" carries with it another assumption, that of "optimality. It implies the question of what is a benefit and how it can be added to other benefits. There are two classic positions on this issue. The "Pareto criterion" states that a net benefit occurs when at least one person is better off and no one is worse off. This, presumably, is the situation in a free market: Each person freely chooses to make the exchange in order to achieve individual benefit. In those cases where the good of many requires that a cost be borne by one (say, in the case of someone's land being required for highway construction), that individual must be compensated for the cost which s/he has to bear. The Pareto criterion is, nonetheless, very restrictive: Income redistribution programs are clearly excluded, and the criterion cannot decide between situations which are "suboptimal" (situations in which costs must be borne by individuals and the only question is which individuals it will be).

The more commonly used position for defining net benefit is the "Kaldor criterion" (Kaldor, 1939). This position states that a net benefit occurs when the sum of the benefits is great enough to offset the costs, whether or not those benefits are used to compensate those who bear the costs. In other words, actual reparation for costs incurred need not be made. Under this criterion, net benefits may be positive although the social benefit may be negative: Robbing the poor to give to the rich may turn out to result in a net benefit. The Kaldor criterion is the basis for benefit/cost analysis.

There is, finally, a third consideration. It is not really an assumption, but more a point to be remembered. Every analysis should consider the "do nothing" option. Every project should be evaluated against the baseline of the present situation. If it is not an improvement over what we have now, don't do it. The point seems obvious, but it is not always acted upon. Doing something, even if it is counterproductive, at least demonstrates that one is aware of the problem and concerned about it. It is difficult to convince people who are in need that the best we can do for them is "nothing."

For Further Information:




© 1996 A.J.Filipovitch
Revised 11 March 2005