PEAK-LOAD PRICING AND OPTIMAL CAPACITY UNDER INDIVISIBILITY CONSTRAINTS

Abstract

The problem of peak-load pricing has been solved at least four times in the postwar literature (by Marcel Boiteux [1], Hendrik Houthakker [4], Peter Steiner [7 ], and Jack Hirshleifer [3 ]), and thus an additional treatment of this subject requires some justification.' Ours is threefold. First, the welfare motivation of these analyses has been generally lacking,2 and it is therefore difficult to evaluate the results obtained. Second, the geometric techniques that have been provided are awkward and unconventional, require that costs be redefined for each change in the number of subperiods, and cannot easily be generalized to handle the case of unequal-duration-subperiod loads. We attempt to remedy each of these deficiencies and, in addition, extend the analysis of peak-load pricing by replacing the usual assumption of fully divisible plant by one in which investment opportunities are discrete. This latter may be of little practical importance,3 but it helps shed additional insights on the welfare attributes of the problem.