risk was substantially reduced, was the system of up-front collection of
rents abandoned, with the consequence of a dramatic increase in the
incomes from taxation, the debt issued, and reduction in the income of
the officials [Weir 1989, Bordo and White 1991 and Brewer 1988].
In the eighteenth century in France, in contrast, the default risk and
monitoring cost were still very high [Homer 1963 and Kiser and Kane
2001], the public finances poor [Brewer 1988 and Ekelund and Tollison
1981], and oligarchs could “cream off huge fortunes at the expense of the
monarch and the public” [Brewer, 1988 p. 130]. At this time, privileges
and perquisites were at least four times higher in France than in England
[Brewer 1988, see also Kiser and Kane 2001]. Even after the French
Revolution, French officials had to pay a sum ranging from three to five
times their salary to buy their job [Zeldin 1973].29
We finally examine the model’s implications regarding the probabil-
ities of monitoring and theft against historical evidence. It follows from
dμ = 1 > 0 and dγ = 2dθa-θ < 0 that an increase in the monitoring cost
∂c t ∂c ∂c 4tb
increases the equilibrium probability of theft and, because the variable
fee is reduced, leads the CA to monitor with a lower probability. Simi-
larly, as dγ = 2dθa-θ < 0, an increase in the default risk will reduce the
∂δ ∂δ 4tb
probability of monitoring. It will, however, not affect the probability of
theft.
As the theory predicts, the rate of theft and of level of monitoring
were remarkable low in England during the eighteenth century [Swart
29If indirect taxation was cumbersome in France, one may ask why the French
monarchs not turned to direct taxation. To some extend they did but they faced re-
strictions there as well. Many of the rich were for historical reasons exempted. There
were also political restrictions because direct taxation was considered particularly
unfair. Moreover, the collection of direct taxes was cumbersome as well because the
valuation of land and properties was difficult.
22