This auction mechanism is, as long as n does not go to infinity, inef-
ficient from the CA’s perspective because the interest rates are private
information. When a potential agent selects a bid below his true value
of the office, he can win and the rent, if won, is increasing from zero to
a positive amount.
III.D. Stage One: The Choice of System of Rent
Extraction
We now analyze the CA’s choice of rent-extraction system, which we
interpret to be reflected by the choice of θ. In a system of direct collection
by government officials, they first extract the rents and then provide
them to the CA. In terms of our model, this system is mimicked by a
high θ , as it generates large ex-post revenues to the CA and small rents
on part of the agent who therefore pays the CA small up-front payments.
When monopoly rights are granted to private contractors, in contrast,
they typically pay the CA before, rather than after, they have extracted
the rents. A low θ reflects this system since the agent’s profit, and hence
up-front payment, is large whereas the CA’s ex-post income is low.
The CA maximizes his disposable incomes, or in other words, the net
present value of its two-period flow of incomes. It collects the winning bid
for the office, ^-1 (a4θ) , up front and borrows as much as possible. The
CA is also assumed to be credit constrained, which restricts it to borrow
(1-c)θ (α-θ)
the net present value of its ex post collection of fees, i.e., t1+δ 2b . We
initially assume that the CA can commit to its ex-ante announcement
of the fee.
When selecting the optimal system of rent extraction, the CA there-
fore has the following problem:
18