2
mitigates the effects of the law of 1/n. In sum, institutional details seem consequential for the
operation of the WSJ proposition.
While existing studies offer important empirical and theoretical refinements of the
law of 1/n, we find that there is still room for improvement on both fronts. At the empirical
level, previous works have used as dependent variables indicators which are far from what
scholars or politicians have in mind when talking about particularistic projects, or “pork”.
Particularistic spending encompasses “projects, programs, and grants that concentrate
benefits in geographically specific constituencies, while spreading their costs across all
constituencies through generalized taxation” (WSJ, 643). Which government programs can
be geographically targeted to a small, well-defined area, perhaps no larger than one district?
Transfer programs, with beneficiaries decided upon as a general class, are almost impossible
to target geographically -- at least not while maintaining the guise of truly universal public
expenditure. Similarly, consumption-based government spending programs are difficult to
target geographically, as small consumption goods (e.g. pencils) are unlikely to be produced
only in one small geographic area, and it is even less likely for their procurement to produce a
visible “landmark” for legislators to display, come election time. In order to be
geographically targeted, a project must involve some physical good that can be localized in a
place.
For this reason, we believe capital expenditures provide a good proxy of
particularistic spending.
5
Capital expenditures are often large physical infrastructure projects,
such as buildings, bridges, and roads, which are visibly located in a place and usually
impossible to relocate or withdraw.
6
As discussed by WSJ, such expenditures benefit
districts both through the consumers’ surplus generated to users of the project (net of their
6
The fact that fixed capital assets cannot be easily withdrawn makes it very different from wages for patronage
jobs, which can be cancelled: it’s easier to fire a worker than it is to remove a dam. This makes fixed assets a
more preferable good from the voters’ perspective.