Returns to Scale From Random Factor Services: Existence and Scope

This note demonstrates that random delivery of services of factors of production is a ubiquitous source of increasing returns to scale with otherwise constant returns technology. It is also shown that such economies must eventually be exhausted -- that returns to scale are asymptotically constant. This creates the possibility that such economies may be empirically unimportant because, in most interesting cases, they are (nearly) exhausted at rather small scales of production. Computation are presented which suggest that this is indeed the case. The model of this note is simpler and more general than the "repairman problem" which exhibits similar properties.