Public Goods 473Identifying a highway of optimal size turns on a comparison of marginal
benefit and marginal cost. The key point to recognize is that the total marginal
benefit to the groups together is found by taking the vertical sum of the separate marginal
benefit (demand) curves.For instance, according to Figure 11.3, a 10-mile-long
highway delivers a marginal benefit of $1.75 million per mile to commercial
vehicles and $1 million per mile to ordinary drivers. Since these trips are non-
rival (i.e., the highway has more than enough capacity for both groups), the
total marginal benefit is $2.75 million. More generally, the uppermost
“demand” curve shows the sum of the groups’ marginal benefits by size of
highway. We can now determine the optimal size of the public project in the
usual way. In the figure, a 17.5-mile highway generates the maximum social net
benefit. At this size, total marginal benefit equals marginal cost.
Two observations are in order. First, there is the problem of financing the
project. As pointed out earlier, to optimize usage (and therefore benefits), the
highway should be toll-free.^17 Consequently, highway costs must be paid
through taxes or government borrowing.
Second, it is difficult to estimate accurately marginal benefits. A sample of
commercial and noncommercial users can be canvassed concerning their
potential usage and value. However, these results are subject to error. The sam-
ple may be unrepresentative, and potential users may deliberately misrepre-
sent their values. Intensive users, eager for the highway to be built (and
knowing it will be collectively financed), have an incentive to overstate their
values. Infrequent users have the incentive to understate their values—to report
zero or even negative values—to block spending on the highway. To the extent
that marginal benefits (and marginal costs) are in error, so, too, will be the
provision of the public good.
Not surprisingly, spending decisions on public goods frequently are
determined as much by politics as by benefit-cost analyses. For instance, the
highway decision could be voted on directly by state representatives. The
virtue of voting is that it is broadly representative of constituents’ preferences.
However, many well-known, unavoidable difficulties are encountered with
systems of voting. Voting often leads to inconsistent results and, in some cir-
cumstances, is subject to undue influence, or even manipulation, by inter-
ested parties—all within its ground rules. In addition, a voter’s ballot, yea or
nay, cannot reflect the magnitude of the individual’s true benefit or cost from
the project. Thus, a project may receive majority approval even though the
dollar gains of the majority fall well short of the total cost incurred by the
minority. Conversely, an economically worthwhile project with benefits dif-
fused over a vast, nonvoting constituency may well be blocked by a special
interest group that gets out its vote.(^17) Governments do set highway tolls to pay back the cost of construction or to raise revenue even
after all borrowing has been retired. To the extent that these tolls reduce usage, they are eco-
nomically inefficient.
c11RegulationPublicGoodsandBenefitCostAnalysis.qxd 9/29/11 1:34 PM Page 473