This paper provides evidence from an empirical test of service contracting
versus bureaucratic production of school bus transportation. Actual contrac
ted cost of school bus service is compared to the estimate of in-house cost
for providing the same service in 19 Tennessee public school systems, in-h
ouse cost is estimated from a two-output, two-input translog cost function
regressed on 91 school system;. Contracted cost is found to be lower than i
n-house cost in IS of the 19 contract systems. The potential savings equals
27% of average contract cost for these 15 systems. In the other 4 contract
systems, actual cost exceeds the estimated in-house cost by 21% of their a
verage contract cost.