Despite the industry's barriers to entry, we constrain UPS's economic moat rating to narrow rather than wide because we expect the firm to outearn its cost of capital by a slim margin--this tempers our confidence that the firm reliably will exceed its cost of capital two decades from now. We consider UPS's air express and ground shipping operations to have economic moats, but its less-than-truckload freight shipping business (the largest LTL operation in the U.S.) earns low margins subject to economic cycles in part because customers have many alternatives, which drives down pricing. Customer switching costs are low and many truckers can provide adequate service.
UPS earns its wide economic moat from efficient scale, cost advantage, and the network effect. Extensive express, ground, and freight networks demand a huge quantity of trucks, trailers, terminals, sorting equipment, IT systems, and skilled labor. Replicating these assets in the absence of sufficient package flow would be costly, and few entities would endure the financial losses during the necessary density-building phase. As evidenced by DHL's worthy effort, such a project would require at least a decade of effort. Even a global shipping powerhouse like DHL failed to displace UPS and FedEx on their massive home turf--these two competitors comprise the efficient scale in high-service U.S. domestic parcel delivery. In this high-fixed-cost business, the substantial parcel volume handled by the incumbents provides a cost advantage that makes competing at market prices difficult for low-volume entrants. Compared with FedEx, UPS produces superior margins via greater package volume, concentration on high-margin ground shipping, and use of a single network rather than parallel air and ground operations. The firm produces attractive ROICs averaging around 15% (excluding the 2007 pension withdrawal expense) despite its intensely asset-based operations. The firm does have substantial asset-light operations in its freight forwarding and contract logistics operations, and the former boast network effects typical of this model--additional offices make the