The BART system was not in our sample. To date it
has not been operating at the level of service for which it
was designed, but our cost functions permit estimating
the opera@ cost from parametric assumptions. The
wage, electricity price, trackage, and output assumptions
are given in Table 3. The estimates were made at a 22.7
million a-mile ammal level of output using Models 3
and 4.
The estimates of BART’s short-run operating costs
(SROC) are not far from the costs actually budgeted in
1975/76. To the extent that BART’s automated design was
intended to be capital intense in order to conserve on labor
expenses, we would expect its operating costs to be lower than our estimates. However, BART is also experiencing
unplanned maintenance and other costs in its start-up
phase which may explain the higher observed expenses.
The cost estimates indicate that BART is operating at
less than optimum long-run capacity at 22.7 million carmiles
annually (short-run total costs are greater than
long-run total costs). Although a point of optimum output
for BART’s endowment of track can be identi@l, (it is
where SRTC = LRTC) it is unlikely that any property
would choose to operate there. That level of output
.would involve extremely long trains running at all times
of day. Using Model 3 and the BART parameters in
Table 3, it appears that the optimum output is abouj 166
million car-miles annually (over seven times the current
output). This “excess” capacity is characteristic of 6xed
guideways and argues for intensive use of the facilities.