In this paper an optimal control model is presented to design enforcem
ent programs minimizing the social costs from both the market and crac
kdown. The model is built around a dynamic equation proposed by Caulki
ns(1) in which the development of the number of dealers in a particula
r illicit drug market depends on market sales and police enforcement.
By using the maximum principle we show that, due to the positive feedb
ack effect hypothesized by Kleiman(2), performing an enforcement polic
y that leads to a collapse of the drug market is more likely to be opt
imal when the sales volume depends on the number of dealers. In case o
f a buyers' market, which means that the total of sales completely dep
ends on the number of buyers, the optimal enforcement policy leads to
a saddle-point equilibrium where the enforcement rate is fixed such th
at the number of dealers is kept constant at a positive level. (C) 199
8 John Wiley & Sons, Ltd.